add files

.gitignore

@@ -0,0 +1,18 @@
+samples/
+wandb/
+outputs/
+__pycache__/
+scripts/animate_inter.py
+scripts/gradio_app.py
+*.ipynb
+*.safetensors
+*.ckpt
+.ossutil_checkpoint/
+ossutil_output/
+debugs/
+.vscode
+.env
+models
+!*/models
+.ipynb_checkpoints
+checkpoints

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 TIGER Lab
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,6 +1,6 @@
 ---
 title: ConsistI2V
-emoji: 💻
+emoji: 🎥
 colorFrom: purple
 colorTo: green
 sdk: gradio

app.py

@@ -0,0 +1,307 @@
+import spaces
+import os
+import json
+import torch
+import random
+import requests
+from PIL import Image
+import numpy as np
+
+import gradio as gr
+from datetime import datetime
+
+import torchvision.transforms as T
+
+from diffusers import DDIMScheduler
+from diffusers.utils.import_utils import is_xformers_available
+from consisti2v.pipelines.pipeline_conditional_animation import ConditionalAnimationPipeline
+from consisti2v.utils.util import save_videos_grid
+from omegaconf import OmegaConf
+
+
+sample_idx     = 0
+scheduler_dict = {
+    "DDIM": DDIMScheduler,
+}
+
+css = """
+.toolbutton {
+    margin-buttom: 0em 0em 0em 0em;
+    max-width: 2.5em;
+    min-width: 2.5em !important;
+    height: 2.5em;
+}
+"""
+
+class AnimateController:
+    def __init__(self):
+        
+        # config dirs
+        self.basedir        = os.getcwd()
+        self.savedir        = os.path.join(self.basedir, "samples/Gradio", datetime.now().strftime("%Y-%m-%dT%H-%M-%S"))
+        self.savedir_sample = os.path.join(self.savedir, "sample")
+        os.makedirs(self.savedir, exist_ok=True)
+
+        self.image_resolution = (256, 256)
+        # config models
+        self.pipeline = ConditionalAnimationPipeline.from_pretrained("TIGER-Lab/ConsistI2V", torch_dtype=torch.float16,)
+        self.pipeline.to("cuda")
+
+    def update_textbox_and_save_image(self, input_image, height_slider, width_slider, center_crop):
+        pil_image = Image.fromarray(input_image.astype(np.uint8)).convert("RGB")
+        img_path = os.path.join(self.savedir, "input_image.png")
+        pil_image.save(img_path)
+        self.image_resolution = pil_image.size
+        pil_image = pil_image.resize((width_slider, height_slider))
+        if center_crop:
+            width, height = width_slider, height_slider
+            aspect_ratio = width / height
+            if aspect_ratio > 16 / 10:
+                pil_image = pil_image.crop((int((width - height * 16 / 10) / 2), 0, int((width + height * 16 / 10) / 2), height))
+            elif aspect_ratio < 16 / 10:
+                pil_image = pil_image.crop((0, int((height - width * 10 / 16) / 2), width, int((height + width * 10 / 16) / 2)))
+        return gr.Textbox.update(value=img_path), gr.Image.update(value=np.array(pil_image))
+
+    @spaces.GPU
+    def animate(
+        self,
+        prompt_textbox, 
+        negative_prompt_textbox,
+        input_image_path,
+        sampler_dropdown, 
+        sample_step_slider, 
+        width_slider, 
+        height_slider, 
+        txt_cfg_scale_slider,
+        img_cfg_scale_slider,
+        center_crop,
+        frame_stride,
+        use_frameinit,
+        frame_init_noise_level,
+        seed_textbox
+    ):
+        if self.pipeline is None:
+            raise gr.Error(f"Please select a pretrained pipeline path.")
+        if input_image_path == "":
+            raise gr.Error(f"Please upload an input image.")
+        if (not center_crop) and (width_slider % 8 != 0 or height_slider % 8 != 0):
+            raise gr.Error(f"`height` and `width` have to be divisible by 8 but are {height_slider} and {width_slider}.")
+        if center_crop and (width_slider % 8 != 0 or height_slider % 8 != 0):
+            raise gr.Error(f"`height` and `width` (after cropping) have to be divisible by 8 but are {height_slider} and {width_slider}.")
+
+        if is_xformers_available() and int(torch.__version__.split(".")[0]) < 2: self.pipeline.unet.enable_xformers_memory_efficient_attention()
+
+        if seed_textbox != -1 and seed_textbox != "": torch.manual_seed(int(seed_textbox))
+        else: torch.seed()
+        seed = torch.initial_seed()
+
+        if input_image_path.startswith("http://") or input_image_path.startswith("https://"):
+            first_frame = Image.open(requests.get(input_image_path, stream=True).raw).convert('RGB')
+        else:
+            first_frame = Image.open(input_image_path).convert('RGB')
+        
+        original_width, original_height = first_frame.size
+
+        if not center_crop:
+            img_transform = T.Compose([
+                T.ToTensor(),
+                T.Resize((height_slider, width_slider), antialias=None),
+                T.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+            ])
+        else:
+            aspect_ratio = original_width / original_height
+            crop_aspect_ratio = width_slider / height_slider
+            if aspect_ratio > crop_aspect_ratio:
+                center_crop_width = int(crop_aspect_ratio * original_height)
+                center_crop_height = original_height
+            elif aspect_ratio < crop_aspect_ratio:
+                center_crop_width = original_width
+                center_crop_height = int(original_width / crop_aspect_ratio)
+            else:
+                center_crop_width = original_width
+                center_crop_height = original_height
+            img_transform = T.Compose([
+                T.ToTensor(),
+                T.CenterCrop((center_crop_height, center_crop_width)),
+                T.Resize((height_slider, width_slider), antialias=None),
+                T.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+            ])
+        
+        first_frame = img_transform(first_frame).unsqueeze(0)
+        first_frame = first_frame.to("cuda")
+
+        if use_frameinit:
+            self.pipeline.init_filter(
+                width         = width_slider,
+                height        = height_slider,
+                video_length  = 16,
+                filter_params = OmegaConf.create({'method': 'gaussian', 'd_s': 0.25, 'd_t': 0.25,})
+            )
+
+
+        sample = self.pipeline(
+            prompt_textbox,
+            negative_prompt       = negative_prompt_textbox,
+            first_frames          = first_frame,
+            num_inference_steps   = sample_step_slider,
+            guidance_scale_txt    = txt_cfg_scale_slider,
+            guidance_scale_img    = img_cfg_scale_slider,
+            width                 = width_slider,
+            height                = height_slider,
+            video_length          = 16,
+            noise_sampling_method = "pyoco_mixed",
+            noise_alpha           = 1.0,
+            frame_stride          = frame_stride,
+            use_frameinit         = use_frameinit,
+            frameinit_noise_level = frame_init_noise_level,
+            camera_motion         = None,
+        ).videos
+
+        global sample_idx
+        sample_idx += 1
+        save_sample_path = os.path.join(self.savedir_sample, f"{sample_idx}.mp4")
+        save_videos_grid(sample, save_sample_path, format="mp4")
+    
+        sample_config = {
+            "prompt": prompt_textbox,
+            "n_prompt": negative_prompt_textbox,
+            "first_frame_path": input_image_path,
+            "sampler": sampler_dropdown,
+            "num_inference_steps": sample_step_slider,
+            "guidance_scale_text": txt_cfg_scale_slider,
+            "guidance_scale_image": img_cfg_scale_slider,
+            "width": width_slider,
+            "height": height_slider,
+            "video_length": 8,
+            "seed": seed
+        }
+        json_str = json.dumps(sample_config, indent=4)
+        with open(os.path.join(self.savedir, "logs.json"), "a") as f:
+            f.write(json_str)
+            f.write("\n\n")
+            
+        return gr.Video.update(value=save_sample_path)
+        
+
+controller = AnimateController()
+
+
+def ui():
+    with gr.Blocks(css=css) as demo:
+        gr.Markdown(
+            """
+            # ConsistI2V Text+Image to Video Generation
+            Input image will be used as the first frame of the video. Text prompts will be used to control the output video content.
+            """
+        )
+
+        with gr.Column(variant="panel"):
+            gr.Markdown(
+                """
+                - Input image can be specified using the "Input Image Path/URL" text box (this can be either a local image path or an image URL) or uploaded by clicking or dragging the image to the "Input Image" box. The uploaded image will be temporarily stored in the "samples/Gradio" folder under the project root folder.
+                - Input image can be resized and/or center cropped to a given resolution by adjusting the "Width" and "Height" sliders. It is recommended to use the same resolution as the training resolution (256x256).
+                - After setting the input image path or changed the width/height of the input image, press the "Preview" button to visualize the resized input image.
+                """
+            )
+            
+            with gr.Row():
+                prompt_textbox = gr.Textbox(label="Prompt", lines=2)
+                negative_prompt_textbox = gr.Textbox(label="Negative prompt", lines=2)
+                
+            with gr.Row().style(equal_height=False):
+                with gr.Column():
+                    with gr.Row():
+                        sampler_dropdown   = gr.Dropdown(label="Sampling method", choices=list(scheduler_dict.keys()), value=list(scheduler_dict.keys())[0])
+                        sample_step_slider = gr.Slider(label="Sampling steps", value=50, minimum=10, maximum=250, step=1)
+                    
+                    with gr.Row():
+                        center_crop   = gr.Checkbox(label="Center Crop the Image", value=True)
+                        width_slider  = gr.Slider(label="Width",  value=256, minimum=0, maximum=512, step=64)
+                        height_slider = gr.Slider(label="Height", value=256, minimum=0, maximum=512, step=64)
+                    with gr.Row():
+                        txt_cfg_scale_slider = gr.Slider(label="Text CFG Scale",   value=7.5, minimum=1.0,   maximum=20.0, step=0.5)
+                        img_cfg_scale_slider = gr.Slider(label="Image CFG Scale",  value=1.0, minimum=1.0,   maximum=20.0, step=0.5)
+                        frame_stride         = gr.Slider(label="Frame Stride",     value=3,   minimum=1,     maximum=5,    step=1)
+                    
+                    with gr.Row():
+                        use_frameinit = gr.Checkbox(label="Enable FrameInit", value=True)
+                        frameinit_noise_level = gr.Slider(label="FrameInit Noise Level", value=850, minimum=1, maximum=999, step=1)
+
+                        
+                        seed_textbox = gr.Textbox(label="Seed", value=-1)
+                        seed_button  = gr.Button(value="\U0001F3B2", elem_classes="toolbutton")
+                        seed_button.click(fn=lambda: gr.Textbox.update(value=random.randint(1, 1e8)), inputs=[], outputs=[seed_textbox])
+                    
+                    
+            
+                    generate_button = gr.Button(value="Generate", variant='primary')
+                
+                with gr.Column():
+                    with gr.Row():
+                        input_image_path = gr.Textbox(label="Input Image Path/URL", lines=1, scale=10, info="Press Enter or the Preview button to confirm the input image.")
+                        preview_button = gr.Button(value="Preview")
+                    
+                    with gr.Row():
+                        input_image = gr.Image(label="Input Image", interactive=True)
+                        input_image.upload(fn=controller.update_textbox_and_save_image, inputs=[input_image, height_slider, width_slider, center_crop], outputs=[input_image_path, input_image])
+                        result_video = gr.Video(label="Generated Animation", interactive=False, autoplay=True)
+
+            def update_and_resize_image(input_image_path, height_slider, width_slider, center_crop):
+                if input_image_path.startswith("http://") or input_image_path.startswith("https://"):
+                    pil_image = Image.open(requests.get(input_image_path, stream=True).raw).convert('RGB')
+                else:
+                    pil_image = Image.open(input_image_path).convert('RGB')
+                controller.image_resolution = pil_image.size
+                original_width, original_height = pil_image.size
+                
+                if center_crop:
+                    crop_aspect_ratio = width_slider / height_slider
+                    aspect_ratio = original_width / original_height
+                    if aspect_ratio > crop_aspect_ratio:
+                        new_width = int(crop_aspect_ratio * original_height)
+                        left = (original_width - new_width) / 2
+                        top = 0
+                        right = left + new_width
+                        bottom = original_height
+                        pil_image = pil_image.crop((left, top, right, bottom))
+                    elif aspect_ratio < crop_aspect_ratio:
+                        new_height = int(original_width / crop_aspect_ratio)
+                        top = (original_height - new_height) / 2
+                        left = 0
+                        right = original_width
+                        bottom = top + new_height
+                        pil_image = pil_image.crop((left, top, right, bottom))
+                
+                pil_image = pil_image.resize((width_slider, height_slider))
+                return gr.Image.update(value=np.array(pil_image))
+            
+            preview_button.click(fn=update_and_resize_image, inputs=[input_image_path, height_slider, width_slider, center_crop], outputs=[input_image])
+            input_image_path.submit(fn=update_and_resize_image, inputs=[input_image_path, height_slider, width_slider, center_crop], outputs=[input_image])
+
+            generate_button.click(
+                fn=controller.animate,
+                inputs=[
+                    prompt_textbox,
+                    negative_prompt_textbox,
+                    input_image_path,
+                    sampler_dropdown,
+                    sample_step_slider,
+                    width_slider,
+                    height_slider,
+                    txt_cfg_scale_slider,
+                    img_cfg_scale_slider,
+                    center_crop,
+                    frame_stride,
+                    use_frameinit,
+                    frameinit_noise_level,
+                    seed_textbox,
+                ],
+                outputs=[result_video]
+            )
+            
+    return demo
+
+
+if __name__ == "__main__":
+    demo = ui()
+    demo.launch(share=True)

configs/inference/inference.yaml

@@ -0,0 +1,48 @@
+output_dir: "samples/inference"
+output_name: "i2v"
+
+pretrained_model_path: "TIGER-Lab/ConsistI2V"
+unet_path: null
+unet_ckpt_prefix: "module."
+pipeline_pretrained_path: null
+
+sampling_kwargs:
+  height: 256
+  width: 256
+  n_frames: 16
+  steps: 50
+  ddim_eta: 0.0
+  guidance_scale_txt: 7.5
+  guidance_scale_img: 1.0
+  guidance_rescale: 0.0
+  num_videos_per_prompt: 1
+  frame_stride: 3
+
+unet_additional_kwargs:
+  variant: null
+  n_temp_heads: 8
+  augment_temporal_attention: true
+  temp_pos_embedding: "rotary" # "rotary" or "sinusoidal"
+  first_frame_condition_mode: "concat"
+  use_frame_stride_condition: true
+  noise_sampling_method: "pyoco_mixed" # "vanilla" or "pyoco_mixed" or "pyoco_progressive"
+  noise_alpha: 1.0
+
+noise_scheduler_kwargs:
+  beta_start: 0.00085
+  beta_end: 0.012
+  beta_schedule: "linear"
+  steps_offset: 1
+  clip_sample: false
+  rescale_betas_zero_snr: false     # true if using zero terminal snr
+  timestep_spacing:       "leading" # "trailing" if using zero terminal snr
+  prediction_type:        "epsilon" # "v_prediction" if using zero terminal snr
+
+frameinit_kwargs:
+  enable: true
+  camera_motion: null
+  noise_level: 850
+  filter_params:
+    method: 'gaussian'
+    d_s: 0.25
+    d_t: 0.25

configs/inference/inference_autoregress.yaml

@@ -0,0 +1,49 @@
+output_dir: "samples/inference"
+output_name: "long_video"
+
+pretrained_model_path: "TIGER-Lab/ConsistI2V"
+unet_path: null
+unet_ckpt_prefix: "module."
+pipeline_pretrained_path: null
+
+sampling_kwargs:
+  height: 256
+  width: 256
+  n_frames: 16
+  steps: 50
+  ddim_eta: 0.0
+  guidance_scale_txt: 7.5
+  guidance_scale_img: 1.0
+  guidance_rescale: 0.0
+  num_videos_per_prompt: 1
+  frame_stride: 3
+  autoregress_steps: 3
+
+unet_additional_kwargs:
+  variant: null
+  n_temp_heads: 8
+  augment_temporal_attention: true
+  temp_pos_embedding: "rotary" # "rotary" or "sinusoidal"
+  first_frame_condition_mode: "concat"
+  use_frame_stride_condition: true
+  noise_sampling_method: "pyoco_mixed" # "vanilla" or "pyoco_mixed" or "pyoco_progressive"
+  noise_alpha: 1.0
+
+noise_scheduler_kwargs:
+  beta_start: 0.00085
+  beta_end: 0.012
+  beta_schedule: "linear"
+  steps_offset: 1
+  clip_sample: false
+  rescale_betas_zero_snr: false     # true if using zero terminal snr
+  timestep_spacing:       "leading" # "trailing" if using zero terminal snr
+  prediction_type:        "epsilon" # "v_prediction" if using zero terminal snr
+
+
+frameinit_kwargs:
+  enable: true
+  noise_level: 850
+  filter_params:
+    method: 'gaussian'
+    d_s: 0.25
+    d_t: 0.25

configs/prompts/default.yaml

@@ -0,0 +1,16 @@
+seeds: random
+
+prompts:
+  - "timelapse at the snow land with aurora in the sky."
+  - "fireworks."
+  - "clown fish swimming through the coral reef."
+  - "melting ice cream dripping down the cone."
+
+n_prompts:
+  - ""
+
+path_to_first_frames:
+  - "assets/example/example_01.png"
+  - "assets/example/example_02.png"
+  - "assets/example/example_03.png"
+  - "assets/example/example_04.png"

configs/training/training.yaml

@@ -0,0 +1,92 @@
+output_dir: "checkpoints"
+pretrained_model_path: "stabilityai/stable-diffusion-2-1-base"
+
+noise_scheduler_kwargs:
+  num_train_timesteps:    1000
+  beta_start:             0.00085
+  beta_end:               0.012
+  beta_schedule:          "linear"
+  steps_offset:           1
+  clip_sample:            false
+  rescale_betas_zero_snr: false     # true if using zero terminal snr
+  timestep_spacing:       "leading" # "trailing" if using zero terminal snr
+  prediction_type:        "epsilon" # "v_prediction" if using zero terminal snr
+
+train_data:
+  dataset:             "joint"
+  pexels_config:
+    enable:            false
+    json_path:         null
+    caption_json_path: null
+    video_folder:      null
+  webvid_config:
+    enable:            true
+    json_path:         "/path/to/webvid/annotation"
+    video_folder:      "/path/to/webvid/data"
+  sample_size:       256
+  sample_duration:   null
+  sample_fps:        null
+  sample_stride:     [1, 5]
+  sample_n_frames:   16
+
+validation_data:
+  prompts:
+    - "timelapse at the snow land with aurora in the sky."
+    - "fireworks."
+    - "clown fish swimming through the coral reef."
+    - "melting ice cream dripping down the cone."
+
+  path_to_first_frames:
+    - "assets/example/example_01.jpg"
+    - "assets/example/example_02.jpg"
+    - "assets/example/example_03.jpg"
+    - "assets/example/example_04.jpg"
+
+  num_inference_steps: 50
+  ddim_eta: 0.0
+  guidance_scale_txt: 7.5
+  guidance_scale_img: 1.0
+  guidance_rescale: 0.0
+  frame_stride: 3
+
+trainable_modules:
+  - "all"
+  # - "conv3ds."
+  # - "tempo_attns."
+
+resume_from_checkpoint: null
+
+unet_additional_kwargs:
+  variant: null
+  n_temp_heads: 8
+  augment_temporal_attention: true
+  temp_pos_embedding: "rotary" # "rotary" or "sinusoidal"
+  first_frame_condition_mode: "concat"
+  use_frame_stride_condition: true
+  noise_sampling_method: "pyoco_mixed" # "vanilla" or "pyoco_mixed" or "pyoco_progressive"
+  noise_alpha: 1.0
+
+cfg_random_null_text_ratio: 0.1
+cfg_random_null_img_ratio: 0.1
+
+use_ema: false
+ema_decay: 0.9999
+
+learning_rate:    5.e-5
+train_batch_size: 3
+gradient_accumulation_steps: 1
+max_grad_norm: 0.5
+
+max_train_epoch:      -1
+max_train_steps:      200000
+checkpointing_epochs: -1
+checkpointing_steps:  2000
+validation_steps:     1000
+
+seed: 42
+mixed_precision: "bf16"
+num_workers: 32
+enable_xformers_memory_efficient_attention: true
+
+is_image: false
+is_debug: false

consisti2v/data/dataset.py

@@ -0,0 +1,315 @@
+import os, io, csv, math, random
+import json
+import numpy as np
+from einops import rearrange
+from decord import VideoReader
+
+import torch
+import torchvision.transforms as transforms
+from torch.utils.data.dataset import Dataset
+
+from diffusers.utils import logging
+
+logger = logging.get_logger(__name__)
+
+class WebVid10M(Dataset):
+    def __init__(
+            self,
+            json_path, video_folder=None,
+            sample_size=256, sample_stride=4, sample_n_frames=16,
+            is_image=False,
+            **kwargs,
+        ):
+        logger.info(f"loading annotations from {json_path} ...")
+        with open(json_path, 'rb') as json_file:
+            json_list = list(json_file)
+        self.dataset = [json.loads(json_str) for json_str in json_list]
+        self.length = len(self.dataset)
+        logger.info(f"data scale: {self.length}")
+
+        self.video_folder    = video_folder
+        self.sample_stride   = sample_stride if isinstance(sample_stride, int) else tuple(sample_stride)
+        self.sample_n_frames = sample_n_frames
+        self.is_image        = is_image
+        
+        sample_size = tuple(sample_size) if not isinstance(sample_size, int) else (sample_size, sample_size)
+        self.pixel_transforms = transforms.Compose([
+            transforms.RandomHorizontalFlip(),
+            transforms.Resize(sample_size[0], antialias=None),
+            transforms.CenterCrop(sample_size),
+            transforms.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+        ])
+    
+    def get_batch(self, idx):
+        video_dict = self.dataset[idx]
+        video_relative_path, name = video_dict['file'], video_dict['text']
+        
+        if self.video_folder is not None:
+            if video_relative_path[0] == '/':
+                video_dir = os.path.join(self.video_folder, os.path.basename(video_relative_path))
+            else:
+                video_dir = os.path.join(self.video_folder, video_relative_path)
+        else:
+            video_dir = video_relative_path
+        video_reader = VideoReader(video_dir)
+        video_length = len(video_reader)
+        
+        if not self.is_image:
+            if isinstance(self.sample_stride, int):
+                stride = self.sample_stride
+            elif isinstance(self.sample_stride, tuple):
+                stride = random.randint(self.sample_stride[0], self.sample_stride[1])
+            clip_length = min(video_length, (self.sample_n_frames - 1) * stride + 1)
+            start_idx   = random.randint(0, video_length - clip_length)
+            batch_index = np.linspace(start_idx, start_idx + clip_length - 1, self.sample_n_frames, dtype=int)
+        else:
+            frame_difference = random.randint(2, self.sample_n_frames)
+            clip_length = min(video_length, (frame_difference - 1) * self.sample_stride + 1)
+            start_idx = random.randint(0, video_length - clip_length)
+            batch_index = [start_idx, start_idx + clip_length - 1]
+
+        pixel_values = torch.from_numpy(video_reader.get_batch(batch_index).asnumpy()).permute(0, 3, 1, 2).contiguous()
+        pixel_values = pixel_values / 255.
+        del video_reader
+        
+        return pixel_values, name
+
+    def __len__(self):
+        return self.length
+
+    def __getitem__(self, idx):
+        while True:
+            try:
+                pixel_values, name = self.get_batch(idx)
+                break
+
+            except Exception as e:
+                idx = random.randint(0, self.length-1)
+
+        pixel_values = self.pixel_transforms(pixel_values)
+        sample = dict(pixel_values=pixel_values, text=name)
+        return sample
+
+
+class Pexels(Dataset):
+    def __init__(
+            self,
+            json_path, caption_json_path, video_folder=None,
+            sample_size=256, sample_duration=1, sample_fps=8,
+            is_image=False,
+            **kwargs,
+        ):
+        logger.info(f"loading captions from {caption_json_path} ...")
+        with open(caption_json_path, 'rb') as caption_json_file:
+            caption_json_list = list(caption_json_file)
+        self.caption_dict = {json.loads(json_str)['id']: json.loads(json_str)['text'] for json_str in caption_json_list}
+        
+        logger.info(f"loading annotations from {json_path} ...")
+        with open(json_path, 'rb') as json_file:
+            json_list = list(json_file)
+        dataset = [json.loads(json_str) for json_str in json_list]
+
+        self.dataset = []
+        for data in dataset:
+            data['text'] = self.caption_dict[data['id']]
+            if data['height'] / data['width'] < 0.625:
+                self.dataset.append(data)
+        self.length = len(self.dataset)
+        logger.info(f"data scale: {self.length}")
+
+        self.video_folder    = video_folder
+        self.sample_duration = sample_duration
+        self.sample_fps      = sample_fps
+        self.sample_n_frames = sample_duration * sample_fps
+        self.is_image        = is_image
+        
+        sample_size = tuple(sample_size) if not isinstance(sample_size, int) else (sample_size, sample_size)
+        self.pixel_transforms = transforms.Compose([
+            transforms.RandomHorizontalFlip(),
+            transforms.Resize(sample_size[0], antialias=None),
+            transforms.CenterCrop(sample_size),
+            transforms.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+        ])
+    
+    def get_batch(self, idx):
+        video_dict = self.dataset[idx]
+        video_relative_path, name = video_dict['file'], video_dict['text']
+        fps = video_dict['fps']
+        
+        if self.video_folder is not None:
+            if video_relative_path[0] == '/':
+                video_dir = os.path.join(self.video_folder, os.path.basename(video_relative_path))
+            else:
+                video_dir = os.path.join(self.video_folder, video_relative_path)
+        else:
+            video_dir = video_relative_path
+        video_reader = VideoReader(video_dir)
+        video_length = len(video_reader)
+        
+        if not self.is_image:
+            clip_length = min(video_length, math.ceil(fps * self.sample_duration))
+            start_idx   = random.randint(0, video_length - clip_length)
+            batch_index = np.linspace(start_idx, start_idx + clip_length - 1, self.sample_n_frames, dtype=int)
+        else:
+            frame_difference = random.randint(2, self.sample_n_frames)
+            sample_stride = math.ceil((fps * self.sample_duration) / (self.sample_n_frames - 1) - 1)
+            clip_length = min(video_length, (frame_difference - 1) * sample_stride + 1)
+            start_idx = random.randint(0, video_length - clip_length)
+            batch_index = [start_idx, start_idx + clip_length - 1]
+
+        pixel_values = torch.from_numpy(video_reader.get_batch(batch_index).asnumpy()).permute(0, 3, 1, 2).contiguous()
+        pixel_values = pixel_values / 255.
+        del video_reader
+        
+        return pixel_values, name
+
+    def __len__(self):
+        return self.length
+
+    def __getitem__(self, idx):
+        while True:
+            try:
+                pixel_values, name = self.get_batch(idx)
+                break
+
+            except Exception as e:
+                idx = random.randint(0, self.length-1)
+
+        pixel_values = self.pixel_transforms(pixel_values)
+        sample = dict(pixel_values=pixel_values, text=name)
+        return sample
+
+
+class JointDataset(Dataset):
+    def __init__(
+            self,
+            webvid_config, pexels_config,
+            sample_size=256,
+            sample_duration=None, sample_fps=None, sample_stride=None, sample_n_frames=None,
+            is_image=False,
+            **kwargs,
+        ):
+        assert (sample_duration is None and sample_fps is None) or (sample_duration is not None and sample_fps is not None), "sample_duration and sample_fps should be both None or not None"
+        if sample_duration is not None and sample_fps is not None:
+            assert sample_stride is None, "when sample_duration and sample_fps are not None, sample_stride should be None"
+        if sample_stride is not None:
+            assert sample_fps is None and sample_duration is None, "when sample_stride is not None, sample_duration and sample_fps should be both None"
+
+        self.dataset = []
+
+        if pexels_config.enable:
+            logger.info(f"loading pexels dataset")
+            logger.info(f"loading captions from {pexels_config.caption_json_path} ...")
+            with open(pexels_config.caption_json_path, 'rb') as caption_json_file:
+                caption_json_list = list(caption_json_file)
+            self.caption_dict = {json.loads(json_str)['id']: json.loads(json_str)['text'] for json_str in caption_json_list}
+            
+            logger.info(f"loading annotations from {pexels_config.json_path} ...")
+            with open(pexels_config.json_path, 'rb') as json_file:
+                json_list = list(json_file)
+            dataset = [json.loads(json_str) for json_str in json_list]
+
+            for data in dataset:
+                data['text'] = self.caption_dict[data['id']]
+                data['dataset'] = 'pexels'
+                if data['height'] / data['width'] < 0.625:
+                    self.dataset.append(data)
+        
+        if webvid_config.enable:
+            logger.info(f"loading webvid dataset")
+            logger.info(f"loading annotations from {webvid_config.json_path} ...")
+            with open(webvid_config.json_path, 'rb') as json_file:
+                json_list = list(json_file)
+            dataset = [json.loads(json_str) for json_str in json_list]
+            for data in dataset:
+                data['dataset'] = 'webvid'
+            self.dataset.extend(dataset)
+
+        self.length = len(self.dataset)
+        logger.info(f"data scale: {self.length}")
+
+        self.pexels_folder   = pexels_config.video_folder
+        self.webvid_folder   = webvid_config.video_folder
+        self.sample_duration = sample_duration
+        self.sample_fps      = sample_fps
+        self.sample_n_frames = sample_duration * sample_fps if sample_n_frames is None else sample_n_frames
+        self.sample_stride   = sample_stride if (sample_stride is None) or (sample_stride is not None and isinstance(sample_stride, int)) else tuple(sample_stride)
+        self.is_image        = is_image
+        
+        sample_size = tuple(sample_size) if not isinstance(sample_size, int) else (sample_size, sample_size)
+        self.pixel_transforms = transforms.Compose([
+            transforms.RandomHorizontalFlip(),
+            transforms.Resize(sample_size[0], antialias=None),
+            transforms.CenterCrop(sample_size),
+            transforms.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+        ])
+    
+    def get_batch(self, idx):
+        video_dict = self.dataset[idx]
+        video_relative_path, name = video_dict['file'], video_dict['text']
+
+        if video_dict['dataset'] == 'pexels':
+            video_folder = self.pexels_folder
+        elif video_dict['dataset'] == 'webvid':
+            video_folder = self.webvid_folder
+        else:
+            raise NotImplementedError
+        
+        if video_folder is not None:
+            if video_relative_path[0] == '/':
+                video_dir = os.path.join(video_folder, os.path.basename(video_relative_path))
+            else:
+                video_dir = os.path.join(video_folder, video_relative_path)
+        else:
+            video_dir = video_relative_path
+        video_reader = VideoReader(video_dir)
+        video_length = len(video_reader)
+        
+        stride = None
+        if not self.is_image:
+            if self.sample_duration is not None:
+                fps = video_dict['fps']
+                clip_length = min(video_length, math.ceil(fps * self.sample_duration))
+            elif self.sample_stride is not None:
+                if isinstance(self.sample_stride, int):
+                    stride = self.sample_stride
+                elif isinstance(self.sample_stride, tuple):
+                    stride = random.randint(self.sample_stride[0], self.sample_stride[1])
+                clip_length = min(video_length, (self.sample_n_frames - 1) * stride + 1)
+
+            start_idx   = random.randint(0, video_length - clip_length)
+            batch_index = np.linspace(start_idx, start_idx + clip_length - 1, self.sample_n_frames, dtype=int)
+
+        else:
+            frame_difference = random.randint(2, self.sample_n_frames)
+            if self.sample_duration is not None:
+                fps = video_dict['fps']
+                sample_stride = math.ceil((fps * self.sample_duration) / (self.sample_n_frames - 1) - 1)
+            elif self.sample_stride is not None:
+                sample_stride = self.sample_stride
+            
+            clip_length = min(video_length, (frame_difference - 1) * sample_stride + 1)
+            start_idx = random.randint(0, video_length - clip_length)
+            batch_index = [start_idx, start_idx + clip_length - 1]
+
+        pixel_values = torch.from_numpy(video_reader.get_batch(batch_index).asnumpy()).permute(0, 3, 1, 2).contiguous()
+        pixel_values = pixel_values / 255.
+        del video_reader
+        
+        return pixel_values, name, stride
+
+    def __len__(self):
+        return self.length
+
+    def __getitem__(self, idx):
+        while True:
+            try:
+                pixel_values, name, stride = self.get_batch(idx)
+                break
+
+            except Exception as e:
+                idx = random.randint(0, self.length-1)
+
+        pixel_values = self.pixel_transforms(pixel_values)
+        sample = dict(pixel_values=pixel_values, text=name, stride=stride)
+        return sample

consisti2v/models/rotary_embedding.py

@@ -0,0 +1,280 @@
+from math import pi, log
+
+import torch
+from torch.nn import Module, ModuleList
+from torch.cuda.amp import autocast
+from torch import nn, einsum, broadcast_tensors, Tensor
+
+from einops import rearrange, repeat
+
+from beartype import beartype
+from beartype.typing import Literal, Union, Optional
+
+# helper functions
+
+def exists(val):
+    return val is not None
+
+def default(val, d):
+    return val if exists(val) else d
+
+# broadcat, as tortoise-tts was using it
+
+def broadcat(tensors, dim = -1):
+    broadcasted_tensors = broadcast_tensors(*tensors)
+    return torch.cat(broadcasted_tensors, dim = dim)
+
+# rotary embedding helper functions
+
+def rotate_half(x):
+    x = rearrange(x, '... (d r) -> ... d r', r = 2)
+    x1, x2 = x.unbind(dim = -1)
+    x = torch.stack((-x2, x1), dim = -1)
+    return rearrange(x, '... d r -> ... (d r)')
+
+@autocast(enabled = False)
+def apply_rotary_emb(freqs, t, start_index = 0, scale = 1., seq_dim = -2):
+    if t.ndim == 3:
+        seq_len = t.shape[seq_dim]
+        freqs = freqs[-seq_len:].to(t)
+
+    rot_dim = freqs.shape[-1]
+    end_index = start_index + rot_dim
+
+    assert rot_dim <= t.shape[-1], f'feature dimension {t.shape[-1]} is not of sufficient size to rotate in all the positions {rot_dim}'
+
+    t_left, t, t_right = t[..., :start_index], t[..., start_index:end_index], t[..., end_index:]
+    t = (t * freqs.cos() * scale) + (rotate_half(t) * freqs.sin() * scale)
+    return torch.cat((t_left, t, t_right), dim = -1)
+
+# learned rotation helpers
+
+def apply_learned_rotations(rotations, t, start_index = 0, freq_ranges = None):
+    if exists(freq_ranges):
+        rotations = einsum('..., f -> ... f', rotations, freq_ranges)
+        rotations = rearrange(rotations, '... r f -> ... (r f)')
+
+    rotations = repeat(rotations, '... n -> ... (n r)', r = 2)
+    return apply_rotary_emb(rotations, t, start_index = start_index)
+
+# classes
+
+class RotaryEmbedding(Module):
+    @beartype
+    def __init__(
+        self,
+        dim,
+        custom_freqs: Optional[Tensor] = None,
+        freqs_for: Union[
+            Literal['lang'],
+            Literal['pixel'],
+            Literal['constant']
+        ] = 'lang',
+        theta = 10000,
+        max_freq = 10,
+        num_freqs = 1,
+        learned_freq = False,
+        use_xpos = False,
+        xpos_scale_base = 512,
+        interpolate_factor = 1.,
+        theta_rescale_factor = 1.,
+        seq_before_head_dim = False
+    ):
+        super().__init__()
+        # proposed by reddit user bloc97, to rescale rotary embeddings to longer sequence length without fine-tuning
+        # has some connection to NTK literature
+        # https://www.reddit.com/r/LocalLLaMA/comments/14lz7j5/ntkaware_scaled_rope_allows_llama_models_to_have/
+
+        theta *= theta_rescale_factor ** (dim / (dim - 2))
+
+        self.freqs_for = freqs_for
+
+        if exists(custom_freqs):
+            freqs = custom_freqs
+        elif freqs_for == 'lang':
+            freqs = 1. / (theta ** (torch.arange(0, dim, 2)[:(dim // 2)].float() / dim))
+        elif freqs_for == 'pixel':
+            freqs = torch.linspace(1., max_freq / 2, dim // 2) * pi
+        elif freqs_for == 'constant':
+            freqs = torch.ones(num_freqs).float()
+
+        self.tmp_store('cached_freqs', None)
+        self.tmp_store('cached_scales', None)
+
+        self.freqs = nn.Parameter(freqs, requires_grad = learned_freq)
+
+        self.learned_freq = learned_freq
+
+        # dummy for device
+
+        self.tmp_store('dummy', torch.tensor(0))
+
+        # default sequence dimension
+
+        self.seq_before_head_dim = seq_before_head_dim
+        self.default_seq_dim = -3 if seq_before_head_dim else -2
+
+        # interpolation factors
+
+        assert interpolate_factor >= 1.
+        self.interpolate_factor = interpolate_factor
+
+        # xpos
+
+        self.use_xpos = use_xpos
+        if not use_xpos:
+            self.tmp_store('scale', None)
+            return
+
+        scale = (torch.arange(0, dim, 2) + 0.4 * dim) / (1.4 * dim)
+        self.scale_base = xpos_scale_base
+        self.tmp_store('scale', scale)
+
+    @property
+    def device(self):
+        return self.dummy.device
+
+    def tmp_store(self, key, value):
+        self.register_buffer(key, value, persistent = False)
+
+    def get_seq_pos(self, seq_len, device, dtype, offset = 0):
+        return (torch.arange(seq_len, device = device, dtype = dtype) + offset) / self.interpolate_factor
+
+    def rotate_queries_or_keys(self, t, seq_dim = None, offset = 0, freq_seq_len = None, seq_pos = None):
+        seq_dim = default(seq_dim, self.default_seq_dim)
+
+        assert not self.use_xpos, 'you must use `.rotate_queries_and_keys` method instead and pass in both queries and keys, for length extrapolatable rotary embeddings'
+
+        device, dtype, seq_len = t.device, t.dtype, t.shape[seq_dim]
+
+        if exists(freq_seq_len):
+            assert freq_seq_len >= seq_len
+            seq_len = freq_seq_len
+
+        if seq_pos is None:
+            seq_pos = self.get_seq_pos(seq_len, device = device, dtype = dtype, offset = offset)
+        else:
+            assert seq_pos.shape[0] == seq_len
+
+        freqs = self.forward(seq_pos, seq_len = seq_len, offset = offset)
+
+        if seq_dim == -3:
+            freqs = rearrange(freqs, 'n d -> n 1 d')
+
+        return apply_rotary_emb(freqs, t, seq_dim = seq_dim)
+
+    def rotate_queries_with_cached_keys(self, q, k, seq_dim = None, offset = 0):
+        seq_dim = default(seq_dim, self.default_seq_dim)
+
+        q_len, k_len = q.shape[seq_dim], k.shape[seq_dim]
+        assert q_len <= k_len
+        rotated_q = self.rotate_queries_or_keys(q, seq_dim = seq_dim, freq_seq_len = k_len)
+        rotated_k = self.rotate_queries_or_keys(k, seq_dim = seq_dim)
+
+        rotated_q = rotated_q.type(q.dtype)
+        rotated_k = rotated_k.type(k.dtype)
+
+        return rotated_q, rotated_k
+
+    def rotate_queries_and_keys(self, q, k, seq_dim = None):
+        seq_dim = default(seq_dim, self.default_seq_dim)
+
+        assert self.use_xpos
+        device, dtype, seq_len = q.device, q.dtype, q.shape[seq_dim]
+
+        seq = self.get_seq_pos(seq_len, dtype = dtype, device = device)
+
+        freqs = self.forward(seq, seq_len = seq_len)
+        scale = self.get_scale(seq, seq_len = seq_len).to(dtype)
+
+        if seq_dim == -3:
+            freqs = rearrange(freqs, 'n d -> n 1 d')
+            scale = rearrange(scale, 'n d -> n 1 d')
+
+        rotated_q = apply_rotary_emb(freqs, q, scale = scale, seq_dim = seq_dim)
+        rotated_k = apply_rotary_emb(freqs, k, scale = scale ** -1, seq_dim = seq_dim)
+
+        rotated_q = rotated_q.type(q.dtype)
+        rotated_k = rotated_k.type(k.dtype)
+
+        return rotated_q, rotated_k
+
+    @beartype
+    def get_scale(
+        self,
+        t: Tensor,
+        seq_len: Optional[int] = None,
+        offset = 0
+    ):
+        assert self.use_xpos
+
+        should_cache = exists(seq_len)
+
+        if (
+            should_cache and \
+            exists(self.cached_scales) and \
+            (seq_len + offset) <= self.cached_scales.shape[0]
+        ):
+            return self.cached_scales[offset:(offset + seq_len)]
+
+        scale = 1.
+        if self.use_xpos:
+            power = (t - len(t) // 2) / self.scale_base
+            scale = self.scale ** rearrange(power, 'n -> n 1')
+            scale = torch.cat((scale, scale), dim = -1)
+
+        if should_cache:
+            self.tmp_store('cached_scales', scale)
+
+        return scale
+
+    def get_axial_freqs(self, *dims):
+        Colon = slice(None)
+        all_freqs = []
+
+        for ind, dim in enumerate(dims):
+            if self.freqs_for == 'pixel':
+                pos = torch.linspace(-1, 1, steps = dim, device = self.device)
+            else:
+                pos = torch.arange(dim, device = self.device)
+
+            freqs = self.forward(pos, seq_len = dim)
+
+            all_axis = [None] * len(dims)
+            all_axis[ind] = Colon
+
+            new_axis_slice = (Ellipsis, *all_axis, Colon)
+            all_freqs.append(freqs[new_axis_slice])
+
+        all_freqs = broadcast_tensors(*all_freqs)
+        return torch.cat(all_freqs, dim = -1)
+
+    @autocast(enabled = False)
+    def forward(
+        self,
+        t: Tensor,
+        seq_len = None,
+        offset = 0
+    ):
+        # should_cache = (
+        #     not self.learned_freq and \
+        #     exists(seq_len) and \
+        #     self.freqs_for != 'pixel'
+        # )
+
+        # if (
+        #     should_cache and \
+        #     exists(self.cached_freqs) and \
+        #     (offset + seq_len) <= self.cached_freqs.shape[0]
+        # ):
+        #     return self.cached_freqs[offset:(offset + seq_len)].detach()
+
+        freqs = self.freqs
+
+        freqs = einsum('..., f -> ... f', t.type(freqs.dtype), freqs)
+        freqs = repeat(freqs, '... n -> ... (n r)', r = 2)
+
+        # if should_cache:
+        #     self.tmp_store('cached_freqs', freqs.detach())
+
+        return freqs

consisti2v/models/videoldm_attention.py

@@ -0,0 +1,809 @@
+from importlib import import_module
+from typing import Callable, Optional, Union
+import math
+
+from einops import rearrange, repeat
+
+import torch
+import torch.nn.functional as F
+from torch import nn
+
+from diffusers.utils import deprecate, logging
+from diffusers.utils.import_utils import is_xformers_available
+from diffusers.utils.torch_utils import maybe_allow_in_graph
+from diffusers.models.lora import LoRACompatibleLinear, LoRALinearLayer
+from diffusers.models.attention_processor import (
+    Attention,
+    AttnAddedKVProcessor,
+    AttnAddedKVProcessor2_0,
+    AttnProcessor,
+    AttnProcessor2_0,
+    SpatialNorm,
+    LORA_ATTENTION_PROCESSORS,
+    CustomDiffusionAttnProcessor,
+    CustomDiffusionXFormersAttnProcessor,
+    SlicedAttnAddedKVProcessor,
+    XFormersAttnAddedKVProcessor,
+    LoRAAttnAddedKVProcessor,
+    XFormersAttnProcessor,
+    LoRAXFormersAttnProcessor,
+    LoRAAttnProcessor,
+    LoRAAttnProcessor2_0,
+    SlicedAttnProcessor,
+    AttentionProcessor
+)
+
+from .rotary_embedding import RotaryEmbedding
+
+
+logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
+
+
+if is_xformers_available():
+    import xformers
+    import xformers.ops
+else:
+    xformers = None
+
+@maybe_allow_in_graph
+class ConditionalAttention(nn.Module):
+    r"""
+    A cross attention layer.
+
+    Parameters:
+        query_dim (`int`): The number of channels in the query.
+        cross_attention_dim (`int`, *optional*):
+            The number of channels in the encoder_hidden_states. If not given, defaults to `query_dim`.
+        heads (`int`,  *optional*, defaults to 8): The number of heads to use for multi-head attention.
+        dim_head (`int`,  *optional*, defaults to 64): The number of channels in each head.
+        dropout (`float`, *optional*, defaults to 0.0): The dropout probability to use.
+        bias (`bool`, *optional*, defaults to False):
+            Set to `True` for the query, key, and value linear layers to contain a bias parameter.
+    """
+
+    def __init__(
+        self,
+        query_dim: int,
+        cross_attention_dim: Optional[int] = None,
+        heads: int = 8,
+        dim_head: int = 64,
+        dropout: float = 0.0,
+        bias=False,
+        upcast_attention: bool = False,
+        upcast_softmax: bool = False,
+        cross_attention_norm: Optional[str] = None,
+        cross_attention_norm_num_groups: int = 32,
+        added_kv_proj_dim: Optional[int] = None,
+        norm_num_groups: Optional[int] = None,
+        spatial_norm_dim: Optional[int] = None,
+        out_bias: bool = True,
+        scale_qk: bool = True,
+        only_cross_attention: bool = False,
+        eps: float = 1e-5,
+        rescale_output_factor: float = 1.0,
+        residual_connection: bool = False,
+        _from_deprecated_attn_block=False,
+        processor: Optional["AttnProcessor"] = None,
+    ):
+        super().__init__()
+        self.inner_dim = dim_head * heads
+        self.cross_attention_dim = cross_attention_dim if cross_attention_dim is not None else query_dim
+        self.upcast_attention = upcast_attention
+        self.upcast_softmax = upcast_softmax
+        self.rescale_output_factor = rescale_output_factor
+        self.residual_connection = residual_connection
+        self.dropout = dropout
+
+        # we make use of this private variable to know whether this class is loaded
+        # with an deprecated state dict so that we can convert it on the fly
+        self._from_deprecated_attn_block = _from_deprecated_attn_block
+
+        self.scale_qk = scale_qk
+        self.scale = dim_head**-0.5 if self.scale_qk else 1.0
+
+        self.heads = heads
+        # for slice_size > 0 the attention score computation
+        # is split across the batch axis to save memory
+        # You can set slice_size with `set_attention_slice`
+        self.sliceable_head_dim = heads
+
+        self.added_kv_proj_dim = added_kv_proj_dim
+        self.only_cross_attention = only_cross_attention
+
+        if self.added_kv_proj_dim is None and self.only_cross_attention:
+            raise ValueError(
+                "`only_cross_attention` can only be set to True if `added_kv_proj_dim` is not None. Make sure to set either `only_cross_attention=False` or define `added_kv_proj_dim`."
+            )
+
+        if norm_num_groups is not None:
+            self.group_norm = nn.GroupNorm(num_channels=query_dim, num_groups=norm_num_groups, eps=eps, affine=True)
+        else:
+            self.group_norm = None
+
+        if spatial_norm_dim is not None:
+            self.spatial_norm = SpatialNorm(f_channels=query_dim, zq_channels=spatial_norm_dim)
+        else:
+            self.spatial_norm = None
+
+        if cross_attention_norm is None:
+            self.norm_cross = None
+        elif cross_attention_norm == "layer_norm":
+            self.norm_cross = nn.LayerNorm(self.cross_attention_dim)
+        elif cross_attention_norm == "group_norm":
+            if self.added_kv_proj_dim is not None:
+                # The given `encoder_hidden_states` are initially of shape
+                # (batch_size, seq_len, added_kv_proj_dim) before being projected
+                # to (batch_size, seq_len, cross_attention_dim). The norm is applied
+                # before the projection, so we need to use `added_kv_proj_dim` as
+                # the number of channels for the group norm.
+                norm_cross_num_channels = added_kv_proj_dim
+            else:
+                norm_cross_num_channels = self.cross_attention_dim
+
+            self.norm_cross = nn.GroupNorm(
+                num_channels=norm_cross_num_channels, num_groups=cross_attention_norm_num_groups, eps=1e-5, affine=True
+            )
+        else:
+            raise ValueError(
+                f"unknown cross_attention_norm: {cross_attention_norm}. Should be None, 'layer_norm' or 'group_norm'"
+            )
+
+        self.to_q = LoRACompatibleLinear(query_dim, self.inner_dim, bias=bias)
+
+        if not self.only_cross_attention:
+            # only relevant for the `AddedKVProcessor` classes
+            self.to_k = LoRACompatibleLinear(self.cross_attention_dim, self.inner_dim, bias=bias)
+            self.to_v = LoRACompatibleLinear(self.cross_attention_dim, self.inner_dim, bias=bias)
+        else:
+            self.to_k = None
+            self.to_v = None
+
+        if self.added_kv_proj_dim is not None:
+            self.add_k_proj = LoRACompatibleLinear(added_kv_proj_dim, self.inner_dim)
+            self.add_v_proj = LoRACompatibleLinear(added_kv_proj_dim, self.inner_dim)
+
+        self.to_out = nn.ModuleList([])
+        self.to_out.append(LoRACompatibleLinear(self.inner_dim, query_dim, bias=out_bias))
+        self.to_out.append(nn.Dropout(dropout))
+
+        # set attention processor
+        # We use the AttnProcessor2_0 by default when torch 2.x is used which uses
+        # torch.nn.functional.scaled_dot_product_attention for native Flash/memory_efficient_attention
+        # but only if it has the default `scale` argument. TODO remove scale_qk check when we move to torch 2.1
+        if processor is None:
+            processor = (
+                AttnProcessor2_0() if hasattr(F, "scaled_dot_product_attention") and self.scale_qk else AttnProcessor()
+            )
+        self.set_processor(processor)
+
+    def set_use_memory_efficient_attention_xformers(
+        self, use_memory_efficient_attention_xformers: bool, attention_op: Optional[Callable] = None
+    ):
+        is_lora = hasattr(self, "processor") and isinstance(
+            self.processor,
+            LORA_ATTENTION_PROCESSORS,
+        )
+        is_custom_diffusion = hasattr(self, "processor") and isinstance(
+            self.processor, (CustomDiffusionAttnProcessor, CustomDiffusionXFormersAttnProcessor)
+        )
+        is_added_kv_processor = hasattr(self, "processor") and isinstance(
+            self.processor,
+            (
+                AttnAddedKVProcessor,
+                AttnAddedKVProcessor2_0,
+                SlicedAttnAddedKVProcessor,
+                XFormersAttnAddedKVProcessor,
+                LoRAAttnAddedKVProcessor,
+            ),
+        )
+
+        if use_memory_efficient_attention_xformers:
+            if is_added_kv_processor and (is_lora or is_custom_diffusion):
+                raise NotImplementedError(
+                    f"Memory efficient attention is currently not supported for LoRA or custom diffusion for attention processor type {self.processor}"
+                )
+            if not is_xformers_available():
+                raise ModuleNotFoundError(
+                    (
+                        "Refer to https://github.com/facebookresearch/xformers for more information on how to install"
+                        " xformers"
+                    ),
+                    name="xformers",
+                )
+            elif not torch.cuda.is_available():
+                raise ValueError(
+                    "torch.cuda.is_available() should be True but is False. xformers' memory efficient attention is"
+                    " only available for GPU "
+                )
+            else:
+                try:
+                    # Make sure we can run the memory efficient attention
+                    _ = xformers.ops.memory_efficient_attention(
+                        torch.randn((1, 2, 40), device="cuda"),
+                        torch.randn((1, 2, 40), device="cuda"),
+                        torch.randn((1, 2, 40), device="cuda"),
+                    )
+                except Exception as e:
+                    raise e
+
+            if is_lora:
+                # TODO (sayakpaul): should we throw a warning if someone wants to use the xformers
+                # variant when using PT 2.0 now that we have LoRAAttnProcessor2_0?
+                processor = LoRAXFormersAttnProcessor(
+                    hidden_size=self.processor.hidden_size,
+                    cross_attention_dim=self.processor.cross_attention_dim,
+                    rank=self.processor.rank,
+                    attention_op=attention_op,
+                )
+                processor.load_state_dict(self.processor.state_dict())
+                processor.to(self.processor.to_q_lora.up.weight.device)
+            elif is_custom_diffusion:
+                processor = CustomDiffusionXFormersAttnProcessor(
+                    train_kv=self.processor.train_kv,
+                    train_q_out=self.processor.train_q_out,
+                    hidden_size=self.processor.hidden_size,
+                    cross_attention_dim=self.processor.cross_attention_dim,
+                    attention_op=attention_op,
+                )
+                processor.load_state_dict(self.processor.state_dict())
+                if hasattr(self.processor, "to_k_custom_diffusion"):
+                    processor.to(self.processor.to_k_custom_diffusion.weight.device)
+            elif is_added_kv_processor:
+                # TODO(Patrick, Suraj, William) - currently xformers doesn't work for UnCLIP
+                # which uses this type of cross attention ONLY because the attention mask of format
+                # [0, ..., -10.000, ..., 0, ...,] is not supported
+                # throw warning
+                logger.info(
+                    "Memory efficient attention with `xformers` might currently not work correctly if an attention mask is required for the attention operation."
+                )
+                processor = XFormersAttnAddedKVProcessor(attention_op=attention_op)
+            else:
+                processor = XFormersAttnProcessor(attention_op=attention_op)
+        else:
+            if is_lora:
+                attn_processor_class = (
+                    LoRAAttnProcessor2_0 if hasattr(F, "scaled_dot_product_attention") else LoRAAttnProcessor
+                )
+                processor = attn_processor_class(
+                    hidden_size=self.processor.hidden_size,
+                    cross_attention_dim=self.processor.cross_attention_dim,
+                    rank=self.processor.rank,
+                )
+                processor.load_state_dict(self.processor.state_dict())
+                processor.to(self.processor.to_q_lora.up.weight.device)
+            elif is_custom_diffusion:
+                processor = CustomDiffusionAttnProcessor(
+                    train_kv=self.processor.train_kv,
+                    train_q_out=self.processor.train_q_out,
+                    hidden_size=self.processor.hidden_size,
+                    cross_attention_dim=self.processor.cross_attention_dim,
+                )
+                processor.load_state_dict(self.processor.state_dict())
+                if hasattr(self.processor, "to_k_custom_diffusion"):
+                    processor.to(self.processor.to_k_custom_diffusion.weight.device)
+            else:
+                # set attention processor
+                # We use the AttnProcessor2_0 by default when torch 2.x is used which uses
+                # torch.nn.functional.scaled_dot_product_attention for native Flash/memory_efficient_attention
+                # but only if it has the default `scale` argument. TODO remove scale_qk check when we move to torch 2.1
+                processor = (
+                    AttnProcessor2_0()
+                    if hasattr(F, "scaled_dot_product_attention") and self.scale_qk
+                    else AttnProcessor()
+                )
+
+        self.set_processor(processor)
+
+    def set_attention_slice(self, slice_size):
+        if slice_size is not None and slice_size > self.sliceable_head_dim:
+            raise ValueError(f"slice_size {slice_size} has to be smaller or equal to {self.sliceable_head_dim}.")
+
+        if slice_size is not None and self.added_kv_proj_dim is not None:
+            processor = SlicedAttnAddedKVProcessor(slice_size)
+        elif slice_size is not None:
+            processor = SlicedAttnProcessor(slice_size)
+        elif self.added_kv_proj_dim is not None:
+            processor = AttnAddedKVProcessor()
+        else:
+            # set attention processor
+            # We use the AttnProcessor2_0 by default when torch 2.x is used which uses
+            # torch.nn.functional.scaled_dot_product_attention for native Flash/memory_efficient_attention
+            # but only if it has the default `scale` argument. TODO remove scale_qk check when we move to torch 2.1
+            processor = (
+                AttnProcessor2_0() if hasattr(F, "scaled_dot_product_attention") and self.scale_qk else AttnProcessor()
+            )
+
+        self.set_processor(processor)
+
+    def set_processor(self, processor: "AttnProcessor"):
+        if (
+            hasattr(self, "processor")
+            and not isinstance(processor, LORA_ATTENTION_PROCESSORS)
+            and self.to_q.lora_layer is not None
+        ):
+            deprecate(
+                "set_processor to offload LoRA",
+                "0.26.0",
+                "In detail, removing LoRA layers via calling `set_processor` or `set_default_attn_processor` is deprecated. Please make sure to call `pipe.unload_lora_weights()` instead.",
+            )
+            # (Patrick, Sayak) - this can be deprecated once PEFT LoRA integration is complete
+            # We need to remove all LoRA layers
+            for module in self.modules():
+                if hasattr(module, "set_lora_layer"):
+                    module.set_lora_layer(None)
+
+        # if current processor is in `self._modules` and if passed `processor` is not, we need to
+        # pop `processor` from `self._modules`
+        if (
+            hasattr(self, "processor")
+            and isinstance(self.processor, torch.nn.Module)
+            and not isinstance(processor, torch.nn.Module)
+        ):
+            logger.info(f"You are removing possibly trained weights of {self.processor} with {processor}")
+            self._modules.pop("processor")
+
+        self.processor = processor
+
+    def get_processor(self, return_deprecated_lora: bool = False) -> "AttentionProcessor":
+        if not return_deprecated_lora:
+            return self.processor
+
+        # TODO(Sayak, Patrick). The rest of the function is needed to ensure backwards compatible
+        # serialization format for LoRA Attention Processors. It should be deleted once the integration
+        # with PEFT is completed.
+        is_lora_activated = {
+            name: module.lora_layer is not None
+            for name, module in self.named_modules()
+            if hasattr(module, "lora_layer")
+        }
+
+        # 1. if no layer has a LoRA activated we can return the processor as usual
+        if not any(is_lora_activated.values()):
+            return self.processor
+
+        # If doesn't apply LoRA do `add_k_proj` or `add_v_proj`
+        is_lora_activated.pop("add_k_proj", None)
+        is_lora_activated.pop("add_v_proj", None)
+        # 2. else it is not posssible that only some layers have LoRA activated
+        if not all(is_lora_activated.values()):
+            raise ValueError(
+                f"Make sure that either all layers or no layers have LoRA activated, but have {is_lora_activated}"
+            )
+
+        # 3. And we need to merge the current LoRA layers into the corresponding LoRA attention processor
+        non_lora_processor_cls_name = self.processor.__class__.__name__
+        lora_processor_cls = getattr(import_module(__name__), "LoRA" + non_lora_processor_cls_name)
+
+        hidden_size = self.inner_dim
+
+        # now create a LoRA attention processor from the LoRA layers
+        if lora_processor_cls in [LoRAAttnProcessor, LoRAAttnProcessor2_0, LoRAXFormersAttnProcessor]:
+            kwargs = {
+                "cross_attention_dim": self.cross_attention_dim,
+                "rank": self.to_q.lora_layer.rank,
+                "network_alpha": self.to_q.lora_layer.network_alpha,
+                "q_rank": self.to_q.lora_layer.rank,
+                "q_hidden_size": self.to_q.lora_layer.out_features,
+                "k_rank": self.to_k.lora_layer.rank,
+                "k_hidden_size": self.to_k.lora_layer.out_features,
+                "v_rank": self.to_v.lora_layer.rank,
+                "v_hidden_size": self.to_v.lora_layer.out_features,
+                "out_rank": self.to_out[0].lora_layer.rank,
+                "out_hidden_size": self.to_out[0].lora_layer.out_features,
+            }
+
+            if hasattr(self.processor, "attention_op"):
+                kwargs["attention_op"] = self.prcoessor.attention_op
+
+            lora_processor = lora_processor_cls(hidden_size, **kwargs)
+            lora_processor.to_q_lora.load_state_dict(self.to_q.lora_layer.state_dict())
+            lora_processor.to_k_lora.load_state_dict(self.to_k.lora_layer.state_dict())
+            lora_processor.to_v_lora.load_state_dict(self.to_v.lora_layer.state_dict())
+            lora_processor.to_out_lora.load_state_dict(self.to_out[0].lora_layer.state_dict())
+        elif lora_processor_cls == LoRAAttnAddedKVProcessor:
+            lora_processor = lora_processor_cls(
+                hidden_size,
+                cross_attention_dim=self.add_k_proj.weight.shape[0],
+                rank=self.to_q.lora_layer.rank,
+                network_alpha=self.to_q.lora_layer.network_alpha,
+            )
+            lora_processor.to_q_lora.load_state_dict(self.to_q.lora_layer.state_dict())
+            lora_processor.to_k_lora.load_state_dict(self.to_k.lora_layer.state_dict())
+            lora_processor.to_v_lora.load_state_dict(self.to_v.lora_layer.state_dict())
+            lora_processor.to_out_lora.load_state_dict(self.to_out[0].lora_layer.state_dict())
+
+            # only save if used
+            if self.add_k_proj.lora_layer is not None:
+                lora_processor.add_k_proj_lora.load_state_dict(self.add_k_proj.lora_layer.state_dict())
+                lora_processor.add_v_proj_lora.load_state_dict(self.add_v_proj.lora_layer.state_dict())
+            else:
+                lora_processor.add_k_proj_lora = None
+                lora_processor.add_v_proj_lora = None
+        else:
+            raise ValueError(f"{lora_processor_cls} does not exist.")
+
+        return lora_processor
+
+    def forward(self, hidden_states, encoder_hidden_states=None, attention_mask=None, **cross_attention_kwargs):
+        # The `Attention` class can call different attention processors / attention functions
+        # here we simply pass along all tensors to the selected processor class
+        # For standard processors that are defined here, `**cross_attention_kwargs` is empty
+        return self.processor(
+            self,
+            hidden_states,
+            encoder_hidden_states=encoder_hidden_states,
+            attention_mask=attention_mask,
+            **cross_attention_kwargs,
+        )
+
+    def batch_to_head_dim(self, tensor):
+        head_size = self.heads
+        batch_size, seq_len, dim = tensor.shape
+        tensor = tensor.reshape(batch_size // head_size, head_size, seq_len, dim)
+        tensor = tensor.permute(0, 2, 1, 3).reshape(batch_size // head_size, seq_len, dim * head_size)
+        return tensor
+
+    def head_to_batch_dim(self, tensor, out_dim=3):
+        head_size = self.heads
+        batch_size, seq_len, dim = tensor.shape
+        tensor = tensor.reshape(batch_size, seq_len, head_size, dim // head_size)
+        tensor = tensor.permute(0, 2, 1, 3)
+
+        if out_dim == 3:
+            tensor = tensor.reshape(batch_size * head_size, seq_len, dim // head_size)
+
+        return tensor
+
+    def get_attention_scores(self, query, key, attention_mask=None):
+        dtype = query.dtype
+        if self.upcast_attention:
+            query = query.float()
+            key = key.float()
+
+        if attention_mask is None:
+            baddbmm_input = torch.empty(
+                query.shape[0], query.shape[1], key.shape[1], dtype=query.dtype, device=query.device
+            )
+            beta = 0
+        else:
+            baddbmm_input = attention_mask
+            beta = 1
+
+        attention_scores = torch.baddbmm(
+            baddbmm_input,
+            query,
+            key.transpose(-1, -2),
+            beta=beta,
+            alpha=self.scale,
+        )
+        del baddbmm_input
+
+        if self.upcast_softmax:
+            attention_scores = attention_scores.float()
+
+        attention_probs = attention_scores.softmax(dim=-1)
+        del attention_scores
+
+        attention_probs = attention_probs.to(dtype)
+
+        return attention_probs
+
+    def prepare_attention_mask(self, attention_mask, target_length, batch_size=None, out_dim=3):
+        if batch_size is None:
+            deprecate(
+                "batch_size=None",
+                "0.22.0",
+                (
+                    "Not passing the `batch_size` parameter to `prepare_attention_mask` can lead to incorrect"
+                    " attention mask preparation and is deprecated behavior. Please make sure to pass `batch_size` to"
+                    " `prepare_attention_mask` when preparing the attention_mask."
+                ),
+            )
+            batch_size = 1
+
+        head_size = self.heads
+        if attention_mask is None:
+            return attention_mask
+
+        current_length: int = attention_mask.shape[-1]
+        if current_length != target_length:
+            if attention_mask.device.type == "mps":
+                # HACK: MPS: Does not support padding by greater than dimension of input tensor.
+                # Instead, we can manually construct the padding tensor.
+                padding_shape = (attention_mask.shape[0], attention_mask.shape[1], target_length)
+                padding = torch.zeros(padding_shape, dtype=attention_mask.dtype, device=attention_mask.device)
+                attention_mask = torch.cat([attention_mask, padding], dim=2)
+            else:
+                # TODO: for pipelines such as stable-diffusion, padding cross-attn mask:
+                #       we want to instead pad by (0, remaining_length), where remaining_length is:
+                #       remaining_length: int = target_length - current_length
+                # TODO: re-enable tests/models/test_models_unet_2d_condition.py#test_model_xattn_padding
+                attention_mask = F.pad(attention_mask, (0, target_length), value=0.0)
+
+        if out_dim == 3:
+            if attention_mask.shape[0] < batch_size * head_size:
+                attention_mask = attention_mask.repeat_interleave(head_size, dim=0)
+        elif out_dim == 4:
+            attention_mask = attention_mask.unsqueeze(1)
+            attention_mask = attention_mask.repeat_interleave(head_size, dim=1)
+
+        return attention_mask
+
+    def norm_encoder_hidden_states(self, encoder_hidden_states):
+        assert self.norm_cross is not None, "self.norm_cross must be defined to call self.norm_encoder_hidden_states"
+
+        if isinstance(self.norm_cross, nn.LayerNorm):
+            encoder_hidden_states = self.norm_cross(encoder_hidden_states)
+        elif isinstance(self.norm_cross, nn.GroupNorm):
+            # Group norm norms along the channels dimension and expects
+            # input to be in the shape of (N, C, *). In this case, we want
+            # to norm along the hidden dimension, so we need to move
+            # (batch_size, sequence_length, hidden_size) ->
+            # (batch_size, hidden_size, sequence_length)
+            encoder_hidden_states = encoder_hidden_states.transpose(1, 2)
+            encoder_hidden_states = self.norm_cross(encoder_hidden_states)
+            encoder_hidden_states = encoder_hidden_states.transpose(1, 2)
+        else:
+            assert False
+
+        return encoder_hidden_states
+
+
+class TemporalConditionalAttention(Attention):
+    def __init__(self, n_frames=8, rotary_emb=False, *args, **kwargs):
+        super().__init__(processor=RotaryEmbAttnProcessor2_0() if rotary_emb else None, *args, **kwargs)
+
+        if not rotary_emb:
+            self.pos_enc = PositionalEncoding(self.inner_dim)
+        else:
+            rotary_bias = RelativePositionBias(heads=kwargs['heads'], max_distance=32)
+            self.rotary_bias = rotary_bias
+            self.rotary_emb = RotaryEmbedding(self.inner_dim // 2)
+
+        self.use_rotary_emb = rotary_emb
+        self.n_frames = n_frames
+
+    def forward(
+        self, 
+        hidden_states, 
+        encoder_hidden_states=None,
+        attention_mask=None,
+        adjacent_slices=None,
+        **cross_attention_kwargs):
+
+        key_pos_idx = None
+
+        bt, hw, c = hidden_states.shape
+        hidden_states = rearrange(hidden_states, '(b t) hw c -> b hw t c', t=self.n_frames)
+        if not self.use_rotary_emb:
+            pos_embed = self.pos_enc(self.n_frames)
+            hidden_states = hidden_states + pos_embed
+        hidden_states = rearrange(hidden_states, 'b hw t c -> (b hw) t c')
+
+        if encoder_hidden_states is not None:
+            assert adjacent_slices is None
+            encoder_hidden_states = encoder_hidden_states[::self.n_frames]
+            encoder_hidden_states = repeat(encoder_hidden_states, 'b n c -> (b hw) n c', hw=hw)
+
+        if adjacent_slices is not None:
+            assert encoder_hidden_states is None
+            adjacent_slices = rearrange(adjacent_slices, 'b c h w n -> b (h w) n c')
+            if not self.use_rotary_emb:
+                first_frame_pos_embed = pos_embed[0:1, :]
+                adjacent_slices = adjacent_slices + first_frame_pos_embed
+            else:
+                pos_idx = torch.arange(self.n_frames, device=hidden_states.device, dtype=hidden_states.dtype)
+                first_frame_pos_pad = torch.zeros(adjacent_slices.shape[2], device=hidden_states.device, dtype=hidden_states.dtype)
+                key_pos_idx = torch.cat([pos_idx, first_frame_pos_pad], dim=0)
+            adjacent_slices = rearrange(adjacent_slices, 'b hw n c -> (b hw) n c')
+            encoder_hidden_states = torch.cat([hidden_states, adjacent_slices], dim=1)
+
+        if not self.use_rotary_emb:
+            out = self.processor(
+                self,
+                hidden_states,
+                encoder_hidden_states=encoder_hidden_states,
+                attention_mask=attention_mask,
+                **cross_attention_kwargs,
+            )
+        else:
+            out = self.processor(
+                self,
+                hidden_states,
+                encoder_hidden_states=encoder_hidden_states,
+                attention_mask=attention_mask,
+                key_pos_idx=key_pos_idx,
+                **cross_attention_kwargs,
+            )
+
+        out = rearrange(out, '(b hw) t c -> (b t) hw c', hw=hw)
+
+        return out
+
+    def set_use_memory_efficient_attention_xformers(self, use_memory_efficient_attention_xformers, attention_op=None):
+        if use_memory_efficient_attention_xformers:
+            try:
+                # Make sure we can run the memory efficient attention
+                _ = xformers.ops.memory_efficient_attention(
+                    torch.randn((1, 2, 40), device="cuda"),
+                    torch.randn((1, 2, 40), device="cuda"),
+                    torch.randn((1, 2, 40), device="cuda"),
+                )
+            except Exception as e:
+                raise e
+            processor = XFormersAttnProcessor(attention_op=attention_op)
+        else:
+            processor = (
+                AttnProcessor2_0()
+                if hasattr(F, "scaled_dot_product_attention") and self.scale_qk
+                else AttnProcessor()
+            )
+        self.set_processor(processor)
+
+
+class PositionalEncoding(nn.Module):
+    def __init__(self, dim, max_pos=512):
+        super().__init__()
+
+        pos = torch.arange(max_pos)
+
+        freq = torch.arange(dim//2) / dim
+        freq = (freq * torch.tensor(10000).log()).exp()
+
+        x = rearrange(pos, 'L -> L 1') / freq
+        x = rearrange(x, 'L d -> L d 1')
+
+        pe = torch.cat((x.sin(), x.cos()), dim=-1)
+        self.pe = rearrange(pe, 'L d sc -> L (d sc)')
+
+        self.dummy = nn.Parameter(torch.rand(1))
+
+    def forward(self, length):
+        enc = self.pe[:length]
+        enc = enc.to(self.dummy.device, self.dummy.dtype)
+        return enc
+
+
+# code taken from https://github.com/Vchitect/LaVie/blob/main/base/models/temporal_attention.py
+class RelativePositionBias(nn.Module):
+    def __init__(
+        self,
+        heads=8,
+        num_buckets=32,
+        max_distance=128,
+    ):
+        super().__init__()
+        self.num_buckets = num_buckets
+        self.max_distance = max_distance
+        self.relative_attention_bias = nn.Embedding(num_buckets, heads)
+
+    @staticmethod
+    def _relative_position_bucket(relative_position, num_buckets=32, max_distance=128):
+        ret = 0
+        n = -relative_position
+
+        num_buckets //= 2
+        ret += (n < 0).long() * num_buckets
+        n = torch.abs(n)
+
+        max_exact = num_buckets // 2
+        is_small = n < max_exact
+
+        val_if_large = max_exact + (
+            torch.log(n.float() / max_exact) / math.log(max_distance / max_exact) * (num_buckets - max_exact)
+        ).long()
+        val_if_large = torch.min(val_if_large, torch.full_like(val_if_large, num_buckets - 1))
+
+        ret += torch.where(is_small, n, val_if_large)
+        return ret
+
+    def forward(self, qlen, klen, device, dtype):
+        q_pos = torch.arange(qlen, dtype = torch.long, device = device)
+        k_pos = torch.arange(klen, dtype = torch.long, device = device)
+        rel_pos = rearrange(k_pos, 'j -> 1 j') - rearrange(q_pos, 'i -> i 1')
+        rp_bucket = self._relative_position_bucket(rel_pos, num_buckets = self.num_buckets, max_distance = self.max_distance)
+        values = self.relative_attention_bias(rp_bucket)
+        values = values.to(device, dtype)
+        return rearrange(values, 'i j h -> h i j') # num_heads, num_frames, num_frames
+
+
+class RotaryEmbAttnProcessor2_0:
+    r"""
+    Processor for implementing scaled dot-product attention (enabled by default if you're using PyTorch 2.0).
+    Add rotary embedding support
+    """
+
+    def __init__(self):
+
+        if not hasattr(F, "scaled_dot_product_attention"):
+            raise ImportError("AttnProcessor2_0 requires PyTorch 2.0, to use it, please upgrade PyTorch to 2.0.")
+
+    def __call__(
+        self,
+        attn: Attention,
+        hidden_states,
+        encoder_hidden_states=None,
+        attention_mask=None,
+        temb=None,
+        scale: float = 1.0,
+        key_pos_idx: Optional[torch.Tensor] = None,
+    ):
+        assert attention_mask is None
+        residual = hidden_states
+
+        if attn.spatial_norm is not None:
+            hidden_states = attn.spatial_norm(hidden_states, temb)
+
+        input_ndim = hidden_states.ndim
+
+        if input_ndim == 4:
+            batch_size, channel, height, width = hidden_states.shape
+            hidden_states = hidden_states.view(batch_size, channel, height * width).transpose(1, 2)
+
+        batch_size, sequence_length, _ = (
+            hidden_states.shape if encoder_hidden_states is None else encoder_hidden_states.shape
+        )
+
+        # if attention_mask is not None:
+        #     attention_mask = attn.prepare_attention_mask(attention_mask, sequence_length, batch_size)
+        #     # scaled_dot_product_attention expects attention_mask shape to be
+        #     # (batch, heads, source_length, target_length)
+        #     attention_mask = attention_mask.view(batch_size, attn.heads, -1, attention_mask.shape[-1])
+
+        if attn.group_norm is not None:
+            hidden_states = attn.group_norm(hidden_states.transpose(1, 2)).transpose(1, 2)
+
+        query = attn.to_q(hidden_states, scale=scale)
+
+        if encoder_hidden_states is None:
+            encoder_hidden_states = hidden_states
+        elif attn.norm_cross:
+            encoder_hidden_states = attn.norm_encoder_hidden_states(encoder_hidden_states)
+        
+        qlen = hidden_states.shape[1]
+        klen = encoder_hidden_states.shape[1]
+        # currently only add bias for self attention. Relative distance doesn't make sense for cross attention.
+        # if qlen == klen:
+        #     time_rel_pos_bias = attn.rotary_bias(qlen, klen, device=hidden_states.device, dtype=hidden_states.dtype)
+        #     attention_mask = repeat(time_rel_pos_bias, "h d1 d2 -> b h d1 d2", b=batch_size)
+
+        key = attn.to_k(encoder_hidden_states, scale=scale)
+        value = attn.to_v(encoder_hidden_states, scale=scale)
+
+        query = attn.rotary_emb.rotate_queries_or_keys(query)
+        if qlen == klen:
+            key = attn.rotary_emb.rotate_queries_or_keys(key)
+        elif key_pos_idx is not None:
+            key = attn.rotary_emb.rotate_queries_or_keys(key, seq_pos=key_pos_idx)
+
+        inner_dim = key.shape[-1]
+        head_dim = inner_dim // attn.heads
+
+        query = query.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
+
+        key = key.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
+        value = value.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
+
+        # the output of sdp = (batch, num_heads, seq_len, head_dim)
+        # TODO: add support for attn.scale when we move to Torch 2.1
+        hidden_states = F.scaled_dot_product_attention(
+            query, key, value, attn_mask=attention_mask, dropout_p=0.0, is_causal=False
+        )
+
+        hidden_states = hidden_states.transpose(1, 2).reshape(batch_size, -1, attn.heads * head_dim)
+        hidden_states = hidden_states.to(query.dtype)
+
+        # linear proj
+        hidden_states = attn.to_out[0](hidden_states, scale=scale)
+        # dropout
+        hidden_states = attn.to_out[1](hidden_states)
+
+        if input_ndim == 4:
+            hidden_states = hidden_states.transpose(-1, -2).reshape(batch_size, channel, height, width)
+
+        if attn.residual_connection:
+            hidden_states = hidden_states + residual
+
+        hidden_states = hidden_states / attn.rescale_output_factor
+
+        return hidden_states

consisti2v/models/videoldm_transformer_blocks.py

@@ -0,0 +1,564 @@
+# Modified from https://github.com/huggingface/diffusers/blob/v0.21.0/src/diffusers/models/transformer_2d.py
+from dataclasses import dataclass
+from typing import Any, Dict, Optional
+
+import torch
+import torch.nn.functional as F
+from torch import nn
+
+from einops import rearrange, repeat
+
+from diffusers.configuration_utils import ConfigMixin, register_to_config
+from diffusers.models.embeddings import ImagePositionalEmbeddings
+from diffusers.utils import BaseOutput, deprecate
+from diffusers.models.attention import AdaLayerNorm, AdaLayerNormZero, FeedForward, GatedSelfAttentionDense
+from diffusers.models.embeddings import PatchEmbed
+from diffusers.models.lora import LoRACompatibleConv, LoRACompatibleLinear
+from diffusers.models.modeling_utils import ModelMixin
+from diffusers.models.transformer_2d import Transformer2DModelOutput
+from diffusers.utils.torch_utils import maybe_allow_in_graph
+from diffusers.models.attention_processor import Attention
+from diffusers.models.lora import LoRACompatibleLinear
+
+from .videoldm_attention import ConditionalAttention, TemporalConditionalAttention
+
+
+class Transformer2DConditionModel(ModelMixin, ConfigMixin):
+    @register_to_config
+    def __init__(
+        self,
+        num_attention_heads: int = 16,
+        attention_head_dim: int = 88,
+        in_channels: Optional[int] = None,
+        out_channels: Optional[int] = None,
+        num_layers: int = 1,
+        dropout: float = 0.0,
+        norm_num_groups: int = 32,
+        cross_attention_dim: Optional[int] = None,
+        attention_bias: bool = False,
+        sample_size: Optional[int] = None,
+        num_vector_embeds: Optional[int] = None,
+        patch_size: Optional[int] = None,
+        activation_fn: str = "geglu",
+        num_embeds_ada_norm: Optional[int] = None,
+        use_linear_projection: bool = False,
+        only_cross_attention: bool = False,
+        double_self_attention: bool = False,
+        upcast_attention: bool = False,
+        norm_type: str = "layer_norm",
+        norm_elementwise_affine: bool = True,
+        attention_type: str = "default",
+        # additional
+        n_frames: int = 8,
+        is_temporal: bool = False,
+        augment_temporal_attention: bool = False,
+        rotary_emb=False,
+    ):
+        super().__init__()
+        self.use_linear_projection = use_linear_projection
+        self.num_attention_heads = num_attention_heads
+        self.attention_head_dim = attention_head_dim
+        inner_dim = num_attention_heads * attention_head_dim
+
+        # 1. Transformer2DModel can process both standard continuous images of shape `(batch_size, num_channels, width, height)` as well as quantized image embeddings of shape `(batch_size, num_image_vectors)`
+        # Define whether input is continuous or discrete depending on configuration
+        self.is_input_continuous = (in_channels is not None) and (patch_size is None)
+        self.is_input_vectorized = num_vector_embeds is not None
+        self.is_input_patches = in_channels is not None and patch_size is not None
+
+        if norm_type == "layer_norm" and num_embeds_ada_norm is not None:
+            deprecation_message = (
+                f"The configuration file of this model: {self.__class__} is outdated. `norm_type` is either not set or"
+                " incorrectly set to `'layer_norm'`.Make sure to set `norm_type` to `'ada_norm'` in the config."
+                " Please make sure to update the config accordingly as leaving `norm_type` might led to incorrect"
+                " results in future versions. If you have downloaded this checkpoint from the Hugging Face Hub, it"
+                " would be very nice if you could open a Pull request for the `transformer/config.json` file"
+            )
+            deprecate("norm_type!=num_embeds_ada_norm", "1.0.0", deprecation_message, standard_warn=False)
+            norm_type = "ada_norm"
+
+        if self.is_input_continuous and self.is_input_vectorized:
+            raise ValueError(
+                f"Cannot define both `in_channels`: {in_channels} and `num_vector_embeds`: {num_vector_embeds}. Make"
+                " sure that either `in_channels` or `num_vector_embeds` is None."
+            )
+        elif self.is_input_vectorized and self.is_input_patches:
+            raise ValueError(
+                f"Cannot define both `num_vector_embeds`: {num_vector_embeds} and `patch_size`: {patch_size}. Make"
+                " sure that either `num_vector_embeds` or `num_patches` is None."
+            )
+        elif not self.is_input_continuous and not self.is_input_vectorized and not self.is_input_patches:
+            raise ValueError(
+                f"Has to define `in_channels`: {in_channels}, `num_vector_embeds`: {num_vector_embeds}, or patch_size:"
+                f" {patch_size}. Make sure that `in_channels`, `num_vector_embeds` or `num_patches` is not None."
+            )
+
+        # 2. Define input layers
+        if self.is_input_continuous:
+            self.in_channels = in_channels
+
+            self.norm = torch.nn.GroupNorm(num_groups=norm_num_groups, num_channels=in_channels, eps=1e-6, affine=True)
+            if use_linear_projection:
+                self.proj_in = LoRACompatibleLinear(in_channels, inner_dim)
+            else:
+                self.proj_in = LoRACompatibleConv(in_channels, inner_dim, kernel_size=1, stride=1, padding=0)
+        elif self.is_input_vectorized:
+            assert sample_size is not None, "Transformer2DModel over discrete input must provide sample_size"
+            assert num_vector_embeds is not None, "Transformer2DModel over discrete input must provide num_embed"
+
+            self.height = sample_size
+            self.width = sample_size
+            self.num_vector_embeds = num_vector_embeds
+            self.num_latent_pixels = self.height * self.width
+
+            self.latent_image_embedding = ImagePositionalEmbeddings(
+                num_embed=num_vector_embeds, embed_dim=inner_dim, height=self.height, width=self.width
+            )
+        elif self.is_input_patches:
+            assert sample_size is not None, "Transformer2DModel over patched input must provide sample_size"
+
+            self.height = sample_size
+            self.width = sample_size
+
+            self.patch_size = patch_size
+            self.pos_embed = PatchEmbed(
+                height=sample_size,
+                width=sample_size,
+                patch_size=patch_size,
+                in_channels=in_channels,
+                embed_dim=inner_dim,
+            )
+
+        # 3. Define transformers blocks
+        self.transformer_blocks = nn.ModuleList(
+            [
+                BasicConditionalTransformerBlock(
+                    inner_dim,
+                    num_attention_heads,
+                    attention_head_dim,
+                    dropout=dropout,
+                    cross_attention_dim=cross_attention_dim,
+                    activation_fn=activation_fn,
+                    num_embeds_ada_norm=num_embeds_ada_norm,
+                    attention_bias=attention_bias,
+                    only_cross_attention=only_cross_attention,
+                    double_self_attention=double_self_attention,
+                    upcast_attention=upcast_attention,
+                    norm_type=norm_type,
+                    norm_elementwise_affine=norm_elementwise_affine,
+                    attention_type=attention_type,
+                    # additional
+                    n_frames=n_frames,
+                    is_temporal=is_temporal,
+                    augment_temporal_attention=augment_temporal_attention,
+                    rotary_emb=rotary_emb,
+                )
+                for d in range(num_layers)
+            ]
+        )
+
+        # 4. Define output layers
+        self.out_channels = in_channels if out_channels is None else out_channels
+        if self.is_input_continuous:
+            # TODO: should use out_channels for continuous projections
+            if use_linear_projection:
+                self.proj_out = LoRACompatibleLinear(inner_dim, in_channels)
+            else:
+                self.proj_out = LoRACompatibleConv(inner_dim, in_channels, kernel_size=1, stride=1, padding=0)
+        elif self.is_input_vectorized:
+            self.norm_out = nn.LayerNorm(inner_dim)
+            self.out = nn.Linear(inner_dim, self.num_vector_embeds - 1)
+        elif self.is_input_patches:
+            self.norm_out = nn.LayerNorm(inner_dim, elementwise_affine=False, eps=1e-6)
+            self.proj_out_1 = nn.Linear(inner_dim, 2 * inner_dim)
+            self.proj_out_2 = nn.Linear(inner_dim, patch_size * patch_size * self.out_channels)
+
+        self.alpha = None
+        if is_temporal:
+            self.alpha = nn.Parameter(torch.ones(1))
+
+        self.gradient_checkpointing = False
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        timestep: Optional[torch.LongTensor] = None,
+        class_labels: Optional[torch.LongTensor] = None,
+        cross_attention_kwargs: Dict[str, Any] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.Tensor] = None,
+        return_dict: bool = True,
+        condition_on_first_frame: bool = False,
+    ):
+        input_states = hidden_states
+        input_height, input_width = hidden_states.shape[-2:]
+        # ensure attention_mask is a bias, and give it a singleton query_tokens dimension.
+        #   we may have done this conversion already, e.g. if we came here via UNet2DConditionModel#forward.
+        #   we can tell by counting dims; if ndim == 2: it's a mask rather than a bias.
+        # expects mask of shape:
+        #   [batch, key_tokens]
+        # adds singleton query_tokens dimension:
+        #   [batch,                    1, key_tokens]
+        # this helps to broadcast it as a bias over attention scores, which will be in one of the following shapes:
+        #   [batch,  heads, query_tokens, key_tokens] (e.g. torch sdp attn)
+        #   [batch * heads, query_tokens, key_tokens] (e.g. xformers or classic attn)
+        if attention_mask is not None and attention_mask.ndim == 2:
+            # assume that mask is expressed as:
+            #   (1 = keep,      0 = discard)
+            # convert mask into a bias that can be added to attention scores:
+            #       (keep = +0,     discard = -10000.0)
+            attention_mask = (1 - attention_mask.to(hidden_states.dtype)) * -10000.0
+            attention_mask = attention_mask.unsqueeze(1)
+
+        # convert encoder_attention_mask to a bias the same way we do for attention_mask
+        if encoder_attention_mask is not None and encoder_attention_mask.ndim == 2:
+            encoder_attention_mask = (1 - encoder_attention_mask.to(hidden_states.dtype)) * -10000.0
+            encoder_attention_mask = encoder_attention_mask.unsqueeze(1)
+
+        # Retrieve lora scale.
+        lora_scale = cross_attention_kwargs.get("scale", 1.0) if cross_attention_kwargs is not None else 1.0
+
+        # 1. Input
+        if self.is_input_continuous:
+            batch, _, height, width = hidden_states.shape
+            residual = hidden_states
+
+            hidden_states = self.norm(hidden_states)
+            if not self.use_linear_projection:
+                hidden_states = self.proj_in(hidden_states, lora_scale)
+                inner_dim = hidden_states.shape[1]
+                hidden_states = hidden_states.permute(0, 2, 3, 1).reshape(batch, height * width, inner_dim)
+            else:
+                inner_dim = hidden_states.shape[1]
+                hidden_states = hidden_states.permute(0, 2, 3, 1).reshape(batch, height * width, inner_dim)
+                hidden_states = self.proj_in(hidden_states, scale=lora_scale)
+
+        elif self.is_input_vectorized:
+            hidden_states = self.latent_image_embedding(hidden_states)
+        elif self.is_input_patches:
+            hidden_states = self.pos_embed(hidden_states)
+
+        # 2. Blocks
+        for block in self.transformer_blocks:
+            if self.training and self.gradient_checkpointing:
+                hidden_states = torch.utils.checkpoint.checkpoint(
+                    block,
+                    hidden_states,
+                    attention_mask,
+                    encoder_hidden_states,
+                    encoder_attention_mask,
+                    timestep,
+                    cross_attention_kwargs,
+                    class_labels,
+                    use_reentrant=False,
+                )
+            else:
+                hidden_states = block(
+                    hidden_states,
+                    attention_mask=attention_mask,
+                    encoder_hidden_states=encoder_hidden_states,
+                    encoder_attention_mask=encoder_attention_mask,
+                    timestep=timestep,
+                    cross_attention_kwargs=cross_attention_kwargs,
+                    class_labels=class_labels,
+                    # additional
+                    condition_on_first_frame=condition_on_first_frame,
+                    input_height=input_height,
+                    input_width=input_width,
+                )
+
+        # 3. Output
+        if self.is_input_continuous:
+            if not self.use_linear_projection:
+                hidden_states = hidden_states.reshape(batch, height, width, inner_dim).permute(0, 3, 1, 2).contiguous()
+                hidden_states = self.proj_out(hidden_states, scale=lora_scale)
+            else:
+                hidden_states = self.proj_out(hidden_states, scale=lora_scale)
+                hidden_states = hidden_states.reshape(batch, height, width, inner_dim).permute(0, 3, 1, 2).contiguous()
+
+            output = hidden_states + residual
+        elif self.is_input_vectorized:
+            hidden_states = self.norm_out(hidden_states)
+            logits = self.out(hidden_states)
+            # (batch, self.num_vector_embeds - 1, self.num_latent_pixels)
+            logits = logits.permute(0, 2, 1)
+
+            # log(p(x_0))
+            output = F.log_softmax(logits.double(), dim=1).float()
+        elif self.is_input_patches:
+            # TODO: cleanup!
+            conditioning = self.transformer_blocks[0].norm1.emb(
+                timestep, class_labels, hidden_dtype=hidden_states.dtype
+            )
+            shift, scale = self.proj_out_1(F.silu(conditioning)).chunk(2, dim=1)
+            hidden_states = self.norm_out(hidden_states) * (1 + scale[:, None]) + shift[:, None]
+            hidden_states = self.proj_out_2(hidden_states)
+
+            # unpatchify
+            height = width = int(hidden_states.shape[1] ** 0.5)
+            hidden_states = hidden_states.reshape(
+                shape=(-1, height, width, self.patch_size, self.patch_size, self.out_channels)
+            )
+            hidden_states = torch.einsum("nhwpqc->nchpwq", hidden_states)
+            output = hidden_states.reshape(
+                shape=(-1, self.out_channels, height * self.patch_size, width * self.patch_size)
+            )
+
+        if self.alpha is not None:
+            with torch.no_grad():
+                self.alpha.clamp_(0, 1)
+
+            output = self.alpha * input_states + (1 - self.alpha) * output
+
+        if not return_dict:
+            return (output,)
+
+        return Transformer2DModelOutput(sample=output)
+
+
+@maybe_allow_in_graph
+class BasicConditionalTransformerBlock(nn.Module):
+    """ transformer block with first frame conditioning """
+    def __init__(
+        self,
+        dim: int,
+        num_attention_heads: int,
+        attention_head_dim: int,
+        dropout=0.0,
+        cross_attention_dim: Optional[int] = None,
+        activation_fn: str = "geglu",
+        num_embeds_ada_norm: Optional[int] = None,
+        attention_bias: bool = False,
+        only_cross_attention: bool = False,
+        double_self_attention: bool = False,
+        upcast_attention: bool = False,
+        norm_elementwise_affine: bool = True,
+        norm_type: str = "layer_norm",
+        final_dropout: bool = False,
+        attention_type: str = "default",
+        # additional
+        n_frames: int = 8,
+        is_temporal: bool = False,
+        augment_temporal_attention: bool = False,
+        rotary_emb=False,
+    ):
+        super().__init__()
+        self.n_frames = n_frames
+        self.only_cross_attention = only_cross_attention
+        self.augment_temporal_attention = augment_temporal_attention
+        self.is_temporal = is_temporal
+
+        self.use_ada_layer_norm_zero = (num_embeds_ada_norm is not None) and norm_type == "ada_norm_zero"
+        self.use_ada_layer_norm = (num_embeds_ada_norm is not None) and norm_type == "ada_norm"
+
+        if norm_type in ("ada_norm", "ada_norm_zero") and num_embeds_ada_norm is None:
+            raise ValueError(
+                f"`norm_type` is set to {norm_type}, but `num_embeds_ada_norm` is not defined. Please make sure to"
+                f" define `num_embeds_ada_norm` if setting `norm_type` to {norm_type}."
+            )
+
+        # Define 3 blocks. Each block has its own normalization layer.
+        # 1. Self-Attn
+        if self.use_ada_layer_norm:
+            self.norm1 = AdaLayerNorm(dim, num_embeds_ada_norm)
+        elif self.use_ada_layer_norm_zero:
+            self.norm1 = AdaLayerNormZero(dim, num_embeds_ada_norm)
+        else:
+            self.norm1 = nn.LayerNorm(dim, elementwise_affine=norm_elementwise_affine)
+        
+        if not is_temporal:
+            self.attn1 = ConditionalAttention(
+                query_dim=dim,
+                heads=num_attention_heads,
+                dim_head=attention_head_dim,
+                dropout=dropout,
+                bias=attention_bias,
+                cross_attention_dim=cross_attention_dim if only_cross_attention else None,
+                upcast_attention=upcast_attention,
+            )
+        else:
+            self.attn1 = TemporalConditionalAttention(
+                query_dim=dim,
+                heads=num_attention_heads,
+                dim_head=attention_head_dim,
+                dropout=dropout,
+                bias=attention_bias,
+                cross_attention_dim=cross_attention_dim if only_cross_attention else None,
+                upcast_attention=upcast_attention,
+                # additional
+                n_frames=n_frames,
+                rotary_emb=rotary_emb,
+            )
+
+        # 2. Cross-Attn
+        if cross_attention_dim is not None or double_self_attention:
+            # We currently only use AdaLayerNormZero for self attention where there will only be one attention block.
+            # I.e. the number of returned modulation chunks from AdaLayerZero would not make sense if returned during
+            # the second cross attention block.
+            self.norm2 = (
+                AdaLayerNorm(dim, num_embeds_ada_norm)
+                if self.use_ada_layer_norm
+                else nn.LayerNorm(dim, elementwise_affine=norm_elementwise_affine)
+            )
+            if not is_temporal:
+                self.attn2 = ConditionalAttention(
+                    query_dim=dim,
+                    cross_attention_dim=cross_attention_dim if not double_self_attention else None,
+                    heads=num_attention_heads,
+                    dim_head=attention_head_dim,
+                    dropout=dropout,
+                    bias=attention_bias,
+                    upcast_attention=upcast_attention,
+                )  # is self-attn if encoder_hidden_states is none
+            else:
+                self.attn2 = TemporalConditionalAttention(
+                    query_dim=dim,
+                    cross_attention_dim=cross_attention_dim if not double_self_attention else None,
+                    heads=num_attention_heads,
+                    dim_head=attention_head_dim,
+                    dropout=dropout,
+                    bias=attention_bias,
+                    upcast_attention=upcast_attention,
+                    # additional
+                    n_frames=n_frames,
+                    rotary_emb=rotary_emb,
+                )
+        else:
+            self.norm2 = None
+            self.attn2 = None
+
+        # 3. Feed-forward
+        self.norm3 = nn.LayerNorm(dim, elementwise_affine=norm_elementwise_affine)
+        self.ff = FeedForward(dim, dropout=dropout, activation_fn=activation_fn, final_dropout=final_dropout)
+
+        # 4. Fuser
+        if attention_type == "gated" or attention_type == "gated-text-image":
+            self.fuser = GatedSelfAttentionDense(dim, cross_attention_dim, num_attention_heads, attention_head_dim)
+
+        # let chunk size default to None
+        self._chunk_size = None
+        self._chunk_dim = 0
+
+    def set_chunk_feed_forward(self, chunk_size: Optional[int], dim: int):
+        # Sets chunk feed-forward
+        self._chunk_size = chunk_size
+        self._chunk_dim = dim
+
+    def forward(
+        self,
+        hidden_states: torch.FloatTensor,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.FloatTensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        timestep: Optional[torch.LongTensor] = None,
+        cross_attention_kwargs: Dict[str, Any] = None,
+        class_labels: Optional[torch.LongTensor] = None,
+        condition_on_first_frame: bool = False,
+        input_height: Optional[int] = None,
+        input_width: Optional[int] = None,
+    ):
+        # Notice that normalization is always applied before the real computation in the following blocks.
+        # 0. Self-Attention
+        if self.use_ada_layer_norm:
+            norm_hidden_states = self.norm1(hidden_states, timestep)
+        elif self.use_ada_layer_norm_zero:
+            norm_hidden_states, gate_msa, shift_mlp, scale_mlp, gate_mlp = self.norm1(
+                hidden_states, timestep, class_labels, hidden_dtype=hidden_states.dtype
+            )
+        else:
+            norm_hidden_states = self.norm1(hidden_states)
+
+        # 1. Retrieve lora scale.
+        lora_scale = cross_attention_kwargs.get("scale", 1.0) if cross_attention_kwargs is not None else 1.0
+
+        # 2. Prepare GLIGEN inputs
+        cross_attention_kwargs = cross_attention_kwargs.copy() if cross_attention_kwargs is not None else {}
+        gligen_kwargs = cross_attention_kwargs.pop("gligen", None)
+
+        if condition_on_first_frame:
+            first_frame_hidden_states = rearrange(norm_hidden_states, '(b f) d h -> b f d h', f=self.n_frames)[:, 0, :, :]
+            first_frame_hidden_states = repeat(first_frame_hidden_states, 'b d h -> b f d h', f=self.n_frames)
+            first_frame_hidden_states = rearrange(first_frame_hidden_states, 'b f d h -> (b f) d h')
+            first_frame_concat_hidden_states = torch.cat((norm_hidden_states, first_frame_hidden_states), dim=1)
+            attn_output = self.attn1(
+                norm_hidden_states,
+                encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else first_frame_concat_hidden_states,
+                attention_mask=attention_mask,
+                **cross_attention_kwargs,
+            )
+        elif self.is_temporal and self.augment_temporal_attention:
+            first_frame_hidden_states = rearrange(norm_hidden_states, '(b f) d h -> b f d h', f=self.n_frames)[:, 0, :, :]
+            first_frame_hidden_states = rearrange(first_frame_hidden_states, 'b (h w) c -> b h w c', h=input_height, w=input_width)
+            first_frame_hidden_states = first_frame_hidden_states.permute(0, 3, 1, 2)
+            padded_first_frame = torch.nn.functional.pad(first_frame_hidden_states, (1, 1, 1, 1), "replicate")
+            first_frame_windows = padded_first_frame.unfold(2, 3, 1).unfold(3, 3, 1)
+            mask = torch.tensor([[1, 1, 1], [1, 0, 1], [1, 1, 1]], dtype=torch.bool)
+            adjacent_slices = first_frame_windows[:, :, :, :, mask]
+            attn_output = self.attn1(
+                norm_hidden_states,
+                encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None,
+                attention_mask=attention_mask,
+                adjacent_slices=adjacent_slices,
+                **cross_attention_kwargs,
+            )
+        else:
+            attn_output = self.attn1(
+                norm_hidden_states,
+                encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None,
+                attention_mask=attention_mask,
+                **cross_attention_kwargs,
+            )
+        if self.use_ada_layer_norm_zero:
+            attn_output = gate_msa.unsqueeze(1) * attn_output
+        hidden_states = attn_output + hidden_states
+
+        # 2.5 GLIGEN Control
+        if gligen_kwargs is not None:
+            hidden_states = self.fuser(hidden_states, gligen_kwargs["objs"])
+        # 2.5 ends
+
+        # 3. Cross-Attention
+        if self.attn2 is not None:
+            norm_hidden_states = (
+                self.norm2(hidden_states, timestep) if self.use_ada_layer_norm else self.norm2(hidden_states)
+            )
+
+            attn_output = self.attn2(
+                norm_hidden_states,
+                encoder_hidden_states=encoder_hidden_states,
+                attention_mask=encoder_attention_mask,
+                **cross_attention_kwargs,
+            )
+            hidden_states = attn_output + hidden_states
+
+        # 4. Feed-forward
+        norm_hidden_states = self.norm3(hidden_states)
+
+        if self.use_ada_layer_norm_zero:
+            norm_hidden_states = norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None]
+
+        if self._chunk_size is not None:
+            # "feed_forward_chunk_size" can be used to save memory
+            if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0:
+                raise ValueError(
+                    f"`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`."
+                )
+
+            num_chunks = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size
+            ff_output = torch.cat(
+                [
+                    self.ff(hid_slice, scale=lora_scale)
+                    for hid_slice in norm_hidden_states.chunk(num_chunks, dim=self._chunk_dim)
+                ],
+                dim=self._chunk_dim,
+            )
+        else:
+            ff_output = self.ff(norm_hidden_states, scale=lora_scale)
+
+        if self.use_ada_layer_norm_zero:
+            ff_output = gate_mlp.unsqueeze(1) * ff_output
+
+        hidden_states = ff_output + hidden_states
+
+        return hidden_states

consisti2v/models/videoldm_unet.py

@@ -0,0 +1,1371 @@
+import os
+import re
+from typing import Optional, Tuple, Union, Dict, List, Any
+from einops import rearrange, repeat
+
+import torch
+import torch.nn as nn
+from diffusers.loaders import UNet2DConditionLoadersMixin
+from diffusers.models import ModelMixin
+from diffusers.models.unet_2d_condition import UNet2DConditionOutput
+from diffusers.models.unet_2d_blocks import UNetMidBlock2DCrossAttn, UNetMidBlock2DSimpleCrossAttn
+from diffusers.models.embeddings import (
+    GaussianFourierProjection,
+    ImageHintTimeEmbedding,
+    ImageProjection,
+    ImageTimeEmbedding,
+    PositionNet,
+    TextImageProjection,
+    TextImageTimeEmbedding,
+    TextTimeEmbedding,
+    TimestepEmbedding,
+    Timesteps,
+)
+from diffusers.models.attention_processor import (
+    ADDED_KV_ATTENTION_PROCESSORS,
+    CROSS_ATTENTION_PROCESSORS,
+    AttentionProcessor,
+    AttnAddedKVProcessor,
+    AttnProcessor,
+)
+from diffusers.models.activations import get_activation
+from diffusers.configuration_utils import register_to_config, ConfigMixin
+from diffusers.models.modeling_utils import load_state_dict, load_model_dict_into_meta
+from diffusers.utils import (
+    CONFIG_NAME,
+    DIFFUSERS_CACHE,
+    FLAX_WEIGHTS_NAME,
+    HF_HUB_OFFLINE,
+    SAFETENSORS_WEIGHTS_NAME,
+    WEIGHTS_NAME,
+    _add_variant,
+    _get_model_file,
+    deprecate,
+    is_accelerate_available,
+    is_torch_version,
+    logging,
+)
+from diffusers import __version__
+
+if is_torch_version(">=", "1.9.0"):
+    _LOW_CPU_MEM_USAGE_DEFAULT = True
+else:
+    _LOW_CPU_MEM_USAGE_DEFAULT = False
+
+
+if is_accelerate_available():
+    import accelerate
+    from accelerate.utils import set_module_tensor_to_device
+    from accelerate.utils.versions import is_torch_version
+
+
+
+from .videoldm_unet_blocks import get_down_block, get_up_block, VideoLDMUNetMidBlock2DCrossAttn
+
+logger = logging.get_logger(__name__)
+
+
+class VideoLDMUNet3DConditionModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin):
+    _supports_gradient_checkpointing = True
+    @register_to_config
+    def __init__(
+        self,
+        sample_size: Optional[int] = None,
+        in_channels: int = 4,
+        out_channels: int = 4,
+        center_input_sample: bool = False,
+        flip_sin_to_cos: bool = True,
+        freq_shift: int = 0,
+        down_block_types: Tuple[str] = (
+            "CrossAttnDownBlock2D", # -> VideoLDMDownBlock
+            "CrossAttnDownBlock2D", # -> VideoLDMDownBlock
+            "CrossAttnDownBlock2D", # -> VideoLDMDownBlock
+            "DownBlock2D",
+        ),
+        mid_block_type: Optional[str] = "UNetMidBlock2DCrossAttn",
+        up_block_types: Tuple[str] = (
+            "UpBlock2D",
+            "CrossAttnUpBlock2D", # -> VideoLDMUpBlock
+            "CrossAttnUpBlock2D", # -> VideoLDMUpBlock
+            "CrossAttnUpBlock2D", # -> VideoLDMUpBlock
+        ),
+        only_cross_attention: Union[bool, Tuple[bool]] = False,
+        block_out_channels: Tuple[int] = (320, 640, 1280, 1280),
+        layers_per_block: Union[int, Tuple[int]] = 2,
+        downsample_padding: int = 1,
+        mid_block_scale_factor: float = 1,
+        dropout: float = 0.0,
+        act_fn: str = "silu",
+        norm_num_groups: Optional[int] = 32,
+        norm_eps: float = 1e-5,
+        cross_attention_dim: Union[int, Tuple[int]] = 1280,
+        transformer_layers_per_block: Union[int, Tuple[int]] = 1,
+        encoder_hid_dim: Optional[int] = None,
+        encoder_hid_dim_type: Optional[str] = None,
+        attention_head_dim: Union[int, Tuple[int]] = 8,
+        num_attention_heads: Optional[Union[int, Tuple[int]]] = None,
+        dual_cross_attention: bool = False,
+        use_linear_projection: bool = False,
+        class_embed_type: Optional[str] = None,
+        addition_embed_type: Optional[str] = None,
+        addition_time_embed_dim: Optional[int] = None,
+        num_class_embeds: Optional[int] = None,
+        upcast_attention: bool = False,
+        resnet_time_scale_shift: str = "default",
+        resnet_skip_time_act: bool = False,
+        resnet_out_scale_factor: int = 1.0,
+        time_embedding_type: str = "positional",
+        time_embedding_dim: Optional[int] = None,
+        time_embedding_act_fn: Optional[str] = None,
+        timestep_post_act: Optional[str] = None,
+        time_cond_proj_dim: Optional[int] = None,
+        conv_in_kernel: int = 3,
+        conv_out_kernel: int = 3,
+        projection_class_embeddings_input_dim: Optional[int] = None,
+        attention_type: str = "default",
+        class_embeddings_concat: bool = False,
+        mid_block_only_cross_attention: Optional[bool] = None,
+        cross_attention_norm: Optional[str] = None,
+        addition_embed_type_num_heads=64,
+        # additional
+        use_temporal: bool = True,
+        n_frames: int = 8,
+        n_temp_heads: int = 8,
+        first_frame_condition_mode: str = "none",
+        augment_temporal_attention: bool = False,
+        temp_pos_embedding: str = "sinusoidal",
+        use_frame_stride_condition: bool = False,
+    ):
+        super().__init__()
+
+        rotary_emb = False
+        if temp_pos_embedding == "rotary":
+            # from rotary_embedding_torch import RotaryEmbedding
+            # rotary_emb = RotaryEmbedding(32)
+            # self.rotary_emb = rotary_emb
+            rotary_emb = True
+            self.rotary_emb = rotary_emb
+
+        self.use_temporal = use_temporal
+        self.augment_temporal_attention = augment_temporal_attention
+
+        assert first_frame_condition_mode in ["none", "concat", "conv2d", "input_only"], f"first_frame_condition_mode: {first_frame_condition_mode} must be one of ['none', 'concat', 'conv2d', 'input_only']"
+        self.first_frame_condition_mode = first_frame_condition_mode
+        latent_channels = in_channels
+
+        self.sample_size = sample_size
+
+        if num_attention_heads is not None:
+            raise ValueError(
+                "At the moment it is not possible to define the number of attention heads via `num_attention_heads` because of a naming issue as described in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131. Passing `num_attention_heads` will only be supported in diffusers v0.19."
+            )
+        
+        num_attention_heads = num_attention_heads or attention_head_dim
+
+        # Check inputs
+        if len(down_block_types) != len(up_block_types):
+            raise ValueError(
+                f"Must provide the same number of `down_block_types` as `up_block_types`. `down_block_types`: {down_block_types}. `up_block_types`: {up_block_types}."
+            )
+
+        if len(block_out_channels) != len(down_block_types):
+            raise ValueError(
+                f"Must provide the same number of `block_out_channels` as `down_block_types`. `block_out_channels`: {block_out_channels}. `down_block_types`: {down_block_types}."
+            )
+
+        if not isinstance(only_cross_attention, bool) and len(only_cross_attention) != len(down_block_types):
+            raise ValueError(
+                f"Must provide the same number of `only_cross_attention` as `down_block_types`. `only_cross_attention`: {only_cross_attention}. `down_block_types`: {down_block_types}."
+            )
+
+        if not isinstance(num_attention_heads, int) and len(num_attention_heads) != len(down_block_types):
+            raise ValueError(
+                f"Must provide the same number of `num_attention_heads` as `down_block_types`. `num_attention_heads`: {num_attention_heads}. `down_block_types`: {down_block_types}."
+            )
+
+        if not isinstance(attention_head_dim, int) and len(attention_head_dim) != len(down_block_types):
+            raise ValueError(
+                f"Must provide the same number of `attention_head_dim` as `down_block_types`. `attention_head_dim`: {attention_head_dim}. `down_block_types`: {down_block_types}."
+            )
+
+        if isinstance(cross_attention_dim, list) and len(cross_attention_dim) != len(down_block_types):
+            raise ValueError(
+                f"Must provide the same number of `cross_attention_dim` as `down_block_types`. `cross_attention_dim`: {cross_attention_dim}. `down_block_types`: {down_block_types}."
+            )
+
+        if not isinstance(layers_per_block, int) and len(layers_per_block) != len(down_block_types):
+            raise ValueError(
+                f"Must provide the same number of `layers_per_block` as `down_block_types`. `layers_per_block`: {layers_per_block}. `down_block_types`: {down_block_types}."
+            )
+
+        # input
+        conv_in_padding = (conv_in_kernel - 1) // 2
+        self.conv_in = nn.Conv2d(
+            in_channels, block_out_channels[0], kernel_size=conv_in_kernel, padding=conv_in_padding
+        )
+
+        # time
+        if time_embedding_type == "fourier":
+            time_embed_dim = time_embedding_dim or block_out_channels[0] * 2
+            if time_embed_dim % 2 != 0:
+                raise ValueError(f"`time_embed_dim` should be divisible by 2, but is {time_embed_dim}.")
+            self.time_proj = GaussianFourierProjection(
+                time_embed_dim // 2, set_W_to_weight=False, log=False, flip_sin_to_cos=flip_sin_to_cos
+            )
+            timestep_input_dim = time_embed_dim
+        elif time_embedding_type == "positional":
+            time_embed_dim = time_embedding_dim or block_out_channels[0] * 4
+
+            self.time_proj = Timesteps(block_out_channels[0], flip_sin_to_cos, freq_shift)
+            timestep_input_dim = block_out_channels[0]
+        else:
+            raise ValueError(
+                f"{time_embedding_type} does not exist. Please make sure to use one of `fourier` or `positional`."
+            )
+
+        self.time_embedding = TimestepEmbedding(
+            timestep_input_dim,
+            time_embed_dim,
+            act_fn=act_fn,
+            post_act_fn=timestep_post_act,
+            cond_proj_dim=time_cond_proj_dim,
+        )
+
+        self.use_frame_stride_condition = use_frame_stride_condition
+        if self.use_frame_stride_condition:
+            self.frame_stride_embedding = TimestepEmbedding(
+                timestep_input_dim,
+                time_embed_dim,
+                act_fn=act_fn,
+                post_act_fn=timestep_post_act,
+                cond_proj_dim=time_cond_proj_dim,
+            )
+            # zero init
+            nn.init.zeros_(self.frame_stride_embedding.linear_2.weight)
+            nn.init.zeros_(self.frame_stride_embedding.linear_2.bias)
+
+        if encoder_hid_dim_type is None and encoder_hid_dim is not None:
+            encoder_hid_dim_type = "text_proj"
+            self.register_to_config(encoder_hid_dim_type=encoder_hid_dim_type)
+            logger.info("encoder_hid_dim_type defaults to 'text_proj' as `encoder_hid_dim` is defined.")
+
+        if encoder_hid_dim is None and encoder_hid_dim_type is not None:
+            raise ValueError(
+                f"`encoder_hid_dim` has to be defined when `encoder_hid_dim_type` is set to {encoder_hid_dim_type}."
+            )
+
+        if encoder_hid_dim_type == "text_proj":
+            self.encoder_hid_proj = nn.Linear(encoder_hid_dim, cross_attention_dim)
+        elif encoder_hid_dim_type == "text_image_proj":
+            # image_embed_dim DOESN'T have to be `cross_attention_dim`. To not clutter the __init__ too much
+            # they are set to `cross_attention_dim` here as this is exactly the required dimension for the currently only use
+            # case when `addition_embed_type == "text_image_proj"` (Kadinsky 2.1)`
+            self.encoder_hid_proj = TextImageProjection(
+                text_embed_dim=encoder_hid_dim,
+                image_embed_dim=cross_attention_dim,
+                cross_attention_dim=cross_attention_dim,
+            )
+        elif encoder_hid_dim_type == "image_proj":
+            # Kandinsky 2.2
+            self.encoder_hid_proj = ImageProjection(
+                image_embed_dim=encoder_hid_dim,
+                cross_attention_dim=cross_attention_dim,
+            )
+        elif encoder_hid_dim_type is not None:
+            raise ValueError(
+                f"encoder_hid_dim_type: {encoder_hid_dim_type} must be None, 'text_proj' or 'text_image_proj'."
+            )
+        else:
+            self.encoder_hid_proj = None
+
+        # class embedding
+        if class_embed_type is None and num_class_embeds is not None:
+            self.class_embedding = nn.Embedding(num_class_embeds, time_embed_dim)
+        elif class_embed_type == "timestep":
+            self.class_embedding = TimestepEmbedding(timestep_input_dim, time_embed_dim, act_fn=act_fn)
+        elif class_embed_type == "identity":
+            self.class_embedding = nn.Identity(time_embed_dim, time_embed_dim)
+        elif class_embed_type == "projection":
+            if projection_class_embeddings_input_dim is None:
+                raise ValueError(
+                    "`class_embed_type`: 'projection' requires `projection_class_embeddings_input_dim` be set"
+                )
+            # The projection `class_embed_type` is the same as the timestep `class_embed_type` except
+            # 1. the `class_labels` inputs are not first converted to sinusoidal embeddings
+            # 2. it projects from an arbitrary input dimension.
+            #
+            # Note that `TimestepEmbedding` is quite general, being mainly linear layers and activations.
+            # When used for embedding actual timesteps, the timesteps are first converted to sinusoidal embeddings.
+            # As a result, `TimestepEmbedding` can be passed arbitrary vectors.
+            self.class_embedding = TimestepEmbedding(projection_class_embeddings_input_dim, time_embed_dim)
+        elif class_embed_type == "simple_projection":
+            if projection_class_embeddings_input_dim is None:
+                raise ValueError(
+                    "`class_embed_type`: 'simple_projection' requires `projection_class_embeddings_input_dim` be set"
+                )
+            self.class_embedding = nn.Linear(projection_class_embeddings_input_dim, time_embed_dim)
+        else:
+            self.class_embedding = None
+
+        if addition_embed_type == "text":
+            if encoder_hid_dim is not None:
+                text_time_embedding_from_dim = encoder_hid_dim
+            else:
+                text_time_embedding_from_dim = cross_attention_dim
+
+            self.add_embedding = TextTimeEmbedding(
+                text_time_embedding_from_dim, time_embed_dim, num_heads=addition_embed_type_num_heads
+            )
+        elif addition_embed_type == "text_image":
+            # text_embed_dim and image_embed_dim DON'T have to be `cross_attention_dim`. To not clutter the __init__ too much
+            # they are set to `cross_attention_dim` here as this is exactly the required dimension for the currently only use
+            # case when `addition_embed_type == "text_image"` (Kadinsky 2.1)`
+            self.add_embedding = TextImageTimeEmbedding(
+                text_embed_dim=cross_attention_dim, image_embed_dim=cross_attention_dim, time_embed_dim=time_embed_dim
+            )
+        elif addition_embed_type == "text_time":
+            self.add_time_proj = Timesteps(addition_time_embed_dim, flip_sin_to_cos, freq_shift)
+            self.add_embedding = TimestepEmbedding(projection_class_embeddings_input_dim, time_embed_dim)
+        elif addition_embed_type == "image":
+            # Kandinsky 2.2
+            self.add_embedding = ImageTimeEmbedding(image_embed_dim=encoder_hid_dim, time_embed_dim=time_embed_dim)
+        elif addition_embed_type == "image_hint":
+            # Kandinsky 2.2 ControlNet
+            self.add_embedding = ImageHintTimeEmbedding(image_embed_dim=encoder_hid_dim, time_embed_dim=time_embed_dim)
+        elif addition_embed_type is not None:
+            raise ValueError(f"addition_embed_type: {addition_embed_type} must be None, 'text' or 'text_image'.")
+
+        if time_embedding_act_fn is None:
+            self.time_embed_act = None
+        else:
+            self.time_embed_act = get_activation(time_embedding_act_fn)
+
+        self.down_blocks = nn.ModuleList([])
+        self.up_blocks = nn.ModuleList([])
+
+        if isinstance(only_cross_attention, bool):
+            if mid_block_only_cross_attention is None:
+                mid_block_only_cross_attention = only_cross_attention
+
+            only_cross_attention = [only_cross_attention] * len(down_block_types)
+
+        if mid_block_only_cross_attention is None:
+            mid_block_only_cross_attention = False
+
+        if isinstance(num_attention_heads, int):
+            num_attention_heads = (num_attention_heads,) * len(down_block_types)
+
+        if isinstance(attention_head_dim, int):
+            attention_head_dim = (attention_head_dim,) * len(down_block_types)
+
+        if isinstance(cross_attention_dim, int):
+            cross_attention_dim = (cross_attention_dim,) * len(down_block_types)
+
+        if isinstance(layers_per_block, int):
+            layers_per_block = [layers_per_block] * len(down_block_types)
+
+        if isinstance(transformer_layers_per_block, int):
+            transformer_layers_per_block = [transformer_layers_per_block] * len(down_block_types)
+
+        if class_embeddings_concat:
+            # The time embeddings are concatenated with the class embeddings. The dimension of the
+            # time embeddings passed to the down, middle, and up blocks is twice the dimension of the
+            # regular time embeddings
+            blocks_time_embed_dim = time_embed_dim * 2
+        else:
+            blocks_time_embed_dim = time_embed_dim
+        # down
+        output_channel = block_out_channels[0]
+        for i, down_block_type in enumerate(down_block_types):
+            input_channel = output_channel
+            output_channel = block_out_channels[i]
+            is_final_block = i == len(block_out_channels) - 1
+
+            down_block = get_down_block(
+                down_block_type,
+                num_layers=layers_per_block[i],
+                transformer_layers_per_block=transformer_layers_per_block[i],
+                in_channels=input_channel,
+                out_channels=output_channel,
+                temb_channels=blocks_time_embed_dim,
+                add_downsample=not is_final_block,
+                resnet_eps=norm_eps,
+                resnet_act_fn=act_fn,
+                resnet_groups=norm_num_groups,
+                cross_attention_dim=cross_attention_dim[i],
+                num_attention_heads=num_attention_heads[i],
+                downsample_padding=downsample_padding,
+                dual_cross_attention=dual_cross_attention,
+                use_linear_projection=use_linear_projection,
+                only_cross_attention=only_cross_attention[i],
+                upcast_attention=upcast_attention,
+                resnet_time_scale_shift=resnet_time_scale_shift,
+                attention_type=attention_type,
+                resnet_skip_time_act=resnet_skip_time_act,
+                resnet_out_scale_factor=resnet_out_scale_factor,
+                cross_attention_norm=cross_attention_norm,
+                attention_head_dim=attention_head_dim[i] if attention_head_dim[i] is not None else output_channel,
+                dropout=dropout,
+                # additional
+                use_temporal=use_temporal,
+                augment_temporal_attention=augment_temporal_attention,
+                n_frames=n_frames,
+                n_temp_heads=n_temp_heads,
+                first_frame_condition_mode=first_frame_condition_mode,
+                latent_channels=latent_channels,
+                rotary_emb=rotary_emb,
+            )
+            self.down_blocks.append(down_block)
+
+        # mid
+        if mid_block_type == "UNetMidBlock2DCrossAttn":
+            self.mid_block = VideoLDMUNetMidBlock2DCrossAttn(
+                transformer_layers_per_block=transformer_layers_per_block[-1],
+                in_channels=block_out_channels[-1],
+                temb_channels=blocks_time_embed_dim,
+                dropout=dropout,
+                resnet_eps=norm_eps,
+                resnet_act_fn=act_fn,
+                output_scale_factor=mid_block_scale_factor,
+                resnet_time_scale_shift=resnet_time_scale_shift,
+                cross_attention_dim=cross_attention_dim[-1],
+                num_attention_heads=num_attention_heads[-1],
+                resnet_groups=norm_num_groups,
+                dual_cross_attention=dual_cross_attention,
+                use_linear_projection=use_linear_projection,
+                upcast_attention=upcast_attention,
+                attention_type=attention_type,
+                # additional
+                use_temporal=use_temporal,
+                n_frames=n_frames,
+                first_frame_condition_mode=first_frame_condition_mode,
+                latent_channels=latent_channels,
+            )
+        elif mid_block_type == "UNetMidBlock2DSimpleCrossAttn":
+            self.mid_block = UNetMidBlock2DSimpleCrossAttn(
+                in_channels=block_out_channels[-1],
+                temb_channels=blocks_time_embed_dim,
+                dropout=dropout,
+                resnet_eps=norm_eps,
+                resnet_act_fn=act_fn,
+                output_scale_factor=mid_block_scale_factor,
+                cross_attention_dim=cross_attention_dim[-1],
+                attention_head_dim=attention_head_dim[-1],
+                resnet_groups=norm_num_groups,
+                resnet_time_scale_shift=resnet_time_scale_shift,
+                skip_time_act=resnet_skip_time_act,
+                only_cross_attention=mid_block_only_cross_attention,
+                cross_attention_norm=cross_attention_norm,
+            )
+        elif mid_block_type is None:
+            self.mid_block = None
+        else:
+            raise ValueError(f"unknown mid_block_type : {mid_block_type}")
+
+        # count how many layers upsample the images
+        self.num_upsamplers = 0
+
+        # up
+        reversed_block_out_channels = list(reversed(block_out_channels))
+        reversed_num_attention_heads = list(reversed(num_attention_heads))
+        reversed_layers_per_block = list(reversed(layers_per_block))
+        reversed_cross_attention_dim = list(reversed(cross_attention_dim))
+        reversed_transformer_layers_per_block = list(reversed(transformer_layers_per_block))
+        only_cross_attention = list(reversed(only_cross_attention))
+
+        output_channel = reversed_block_out_channels[0]
+        for i, up_block_type in enumerate(up_block_types):
+            is_final_block = i == len(block_out_channels) - 1
+
+            prev_output_channel = output_channel
+            output_channel = reversed_block_out_channels[i]
+            input_channel = reversed_block_out_channels[min(i + 1, len(block_out_channels) - 1)]
+
+            # add upsample block for all BUT final layer
+            if not is_final_block:
+                add_upsample = True
+                self.num_upsamplers += 1
+            else:
+                add_upsample = False
+
+            up_block = get_up_block(
+                up_block_type,
+                num_layers=reversed_layers_per_block[i] + 1,
+                transformer_layers_per_block=reversed_transformer_layers_per_block[i],
+                in_channels=input_channel,
+                out_channels=output_channel,
+                prev_output_channel=prev_output_channel,
+                temb_channels=blocks_time_embed_dim,
+                add_upsample=add_upsample,
+                resnet_eps=norm_eps,
+                resnet_act_fn=act_fn,
+                resnet_groups=norm_num_groups,
+                cross_attention_dim=reversed_cross_attention_dim[i],
+                num_attention_heads=reversed_num_attention_heads[i],
+                dual_cross_attention=dual_cross_attention,
+                use_linear_projection=use_linear_projection,
+                only_cross_attention=only_cross_attention[i],
+                upcast_attention=upcast_attention,
+                resnet_time_scale_shift=resnet_time_scale_shift,
+                attention_type=attention_type,
+                resnet_skip_time_act=resnet_skip_time_act,
+                resnet_out_scale_factor=resnet_out_scale_factor,
+                cross_attention_norm=cross_attention_norm,
+                attention_head_dim=attention_head_dim[i] if attention_head_dim[i] is not None else output_channel,
+                dropout=dropout,
+                # additional
+                use_temporal=use_temporal,
+                augment_temporal_attention=augment_temporal_attention,
+                n_frames=n_frames,
+                n_temp_heads=n_temp_heads,
+                first_frame_condition_mode=first_frame_condition_mode,
+                latent_channels=latent_channels,
+                rotary_emb=rotary_emb,
+            )
+            self.up_blocks.append(up_block)
+            prev_output_channel = output_channel
+
+        # out
+        if norm_num_groups is not None:
+            self.conv_norm_out = nn.GroupNorm(
+                num_channels=block_out_channels[0], num_groups=norm_num_groups, eps=norm_eps
+            )
+
+            self.conv_act = get_activation(act_fn)
+
+        else:
+            self.conv_norm_out = None
+            self.conv_act = None
+
+        conv_out_padding = (conv_out_kernel - 1) // 2
+        self.conv_out = nn.Conv2d(
+            block_out_channels[0], out_channels, kernel_size=conv_out_kernel, padding=conv_out_padding
+        )
+    
+    @property
+    def attn_processors(self) -> Dict[str, AttentionProcessor]:
+        r"""
+        Returns:
+            `dict` of attention processors: A dictionary containing all attention processors used in the model with
+            indexed by its weight name.
+        """
+        # set recursively
+        processors = {}
+
+        def fn_recursive_add_processors(name: str, module: torch.nn.Module, processors: Dict[str, AttentionProcessor]):
+            if hasattr(module, "get_processor"):
+                processors[f"{name}.processor"] = module.get_processor(return_deprecated_lora=True)
+
+            for sub_name, child in module.named_children():
+                fn_recursive_add_processors(f"{name}.{sub_name}", child, processors)
+
+            return processors
+
+        for name, module in self.named_children():
+            fn_recursive_add_processors(name, module, processors)
+
+        return processors
+
+    def set_attn_processor(self, processor: Union[AttentionProcessor, Dict[str, AttentionProcessor]]):
+        r"""
+        Sets the attention processor to use to compute attention.
+
+        Parameters:
+            processor (`dict` of `AttentionProcessor` or only `AttentionProcessor`):
+                The instantiated processor class or a dictionary of processor classes that will be set as the processor
+                for **all** `Attention` layers.
+
+                If `processor` is a dict, the key needs to define the path to the corresponding cross attention
+                processor. This is strongly recommended when setting trainable attention processors.
+
+        """
+        count = len(self.attn_processors.keys())
+
+        if isinstance(processor, dict) and len(processor) != count:
+            raise ValueError(
+                f"A dict of processors was passed, but the number of processors {len(processor)} does not match the"
+                f" number of attention layers: {count}. Please make sure to pass {count} processor classes."
+            )
+
+        def fn_recursive_attn_processor(name: str, module: torch.nn.Module, processor):
+            if hasattr(module, "set_processor"):
+                if not isinstance(processor, dict):
+                    module.set_processor(processor)
+                else:
+                    module.set_processor(processor.pop(f"{name}.processor"))
+
+            for sub_name, child in module.named_children():
+                fn_recursive_attn_processor(f"{name}.{sub_name}", child, processor)
+
+        for name, module in self.named_children():
+            fn_recursive_attn_processor(name, module, processor)
+
+    def set_default_attn_processor(self):
+        """
+        Disables custom attention processors and sets the default attention implementation.
+        """
+        if all(proc.__class__ in ADDED_KV_ATTENTION_PROCESSORS for proc in self.attn_processors.values()):
+            processor = AttnAddedKVProcessor()
+        elif all(proc.__class__ in CROSS_ATTENTION_PROCESSORS for proc in self.attn_processors.values()):
+            processor = AttnProcessor()
+        else:
+            raise ValueError(
+                f"Cannot call `set_default_attn_processor` when attention processors are of type {next(iter(self.attn_processors.values()))}"
+            )
+
+        self.set_attn_processor(processor)
+
+    def set_attention_slice(self, slice_size):
+        r"""
+        Enable sliced attention computation.
+
+        When this option is enabled, the attention module splits the input tensor in slices to compute attention in
+        several steps. This is useful for saving some memory in exchange for a small decrease in speed.
+
+        Args:
+            slice_size (`str` or `int` or `list(int)`, *optional*, defaults to `"auto"`):
+                When `"auto"`, input to the attention heads is halved, so attention is computed in two steps. If
+                `"max"`, maximum amount of memory is saved by running only one slice at a time. If a number is
+                provided, uses as many slices as `attention_head_dim // slice_size`. In this case, `attention_head_dim`
+                must be a multiple of `slice_size`.
+        """
+        sliceable_head_dims = []
+
+        def fn_recursive_retrieve_sliceable_dims(module: torch.nn.Module):
+            if hasattr(module, "set_attention_slice"):
+                sliceable_head_dims.append(module.sliceable_head_dim)
+
+            for child in module.children():
+                fn_recursive_retrieve_sliceable_dims(child)
+
+        # retrieve number of attention layers
+        for module in self.children():
+            fn_recursive_retrieve_sliceable_dims(module)
+
+        num_sliceable_layers = len(sliceable_head_dims)
+
+        if slice_size == "auto":
+            # half the attention head size is usually a good trade-off between
+            # speed and memory
+            slice_size = [dim // 2 for dim in sliceable_head_dims]
+        elif slice_size == "max":
+            # make smallest slice possible
+            slice_size = num_sliceable_layers * [1]
+
+        slice_size = num_sliceable_layers * [slice_size] if not isinstance(slice_size, list) else slice_size
+
+        if len(slice_size) != len(sliceable_head_dims):
+            raise ValueError(
+                f"You have provided {len(slice_size)}, but {self.config} has {len(sliceable_head_dims)} different"
+                f" attention layers. Make sure to match `len(slice_size)` to be {len(sliceable_head_dims)}."
+            )
+
+        for i in range(len(slice_size)):
+            size = slice_size[i]
+            dim = sliceable_head_dims[i]
+            if size is not None and size > dim:
+                raise ValueError(f"size {size} has to be smaller or equal to {dim}.")
+
+        # Recursively walk through all the children.
+        # Any children which exposes the set_attention_slice method
+        # gets the message
+        def fn_recursive_set_attention_slice(module: torch.nn.Module, slice_size: List[int]):
+            if hasattr(module, "set_attention_slice"):
+                module.set_attention_slice(slice_size.pop())
+
+            for child in module.children():
+                fn_recursive_set_attention_slice(child, slice_size)
+
+        reversed_slice_size = list(reversed(slice_size))
+        for module in self.children():
+            fn_recursive_set_attention_slice(module, reversed_slice_size)
+
+    def _set_gradient_checkpointing(self, module, value=False):
+        if hasattr(module, "gradient_checkpointing"):
+            module.gradient_checkpointing = value
+
+    def forward(
+        self,
+        sample: torch.FloatTensor,
+        timestep: Union[torch.Tensor, float, int],
+        encoder_hidden_states: torch.Tensor,
+        class_labels: Optional[torch.Tensor] = None,
+        timestep_cond: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
+        added_cond_kwargs: Optional[Dict[str, torch.Tensor]] = None,
+        down_block_additional_residuals: Optional[Tuple[torch.Tensor]] = None,
+        mid_block_additional_residual: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.Tensor] = None,
+        return_dict: bool = True,
+        # additional
+        first_frame_latents: Optional[torch.Tensor] = None,
+        frame_stride: Optional[Union[torch.Tensor, float, int]] = None,
+    ) -> Union[UNet2DConditionOutput, Tuple]:
+        # reshape video data
+        assert sample.dim() == 5, f"Expected hidden_states to have ndim=5, but got ndim={sample.dim()}."
+        video_length = sample.shape[2]
+        
+        if first_frame_latents is not None:
+            assert self.config.first_frame_condition_mode != "none", "first_frame_latents is not None, but first_frame_condition_mode is 'none'."
+        
+        if self.config.first_frame_condition_mode != "none":
+            sample = torch.cat([first_frame_latents, sample], dim=2)
+            video_length += 1
+        
+        # copy conditioning embeddings for cross attention
+        if encoder_hidden_states is not None:
+            encoder_hidden_states = repeat(encoder_hidden_states, 'b n c -> (b f) n c', f=video_length)
+        
+        sample = rearrange(sample, "b c f h w -> (b f) c h w")
+
+        # By default samples have to be AT least a multiple of the overall upsampling factor.
+        # The overall upsampling factor is equal to 2 ** (# num of upsampling layers).
+        # However, the upsampling interpolation output size can be forced to fit any upsampling size
+        # on the fly if necessary.
+        default_overall_up_factor = 2**self.num_upsamplers
+
+        # upsample size should be forwarded when sample is not a multiple of `default_overall_up_factor`
+        forward_upsample_size = False
+        upsample_size = None
+
+        if any(s % default_overall_up_factor != 0 for s in sample.shape[-2:]):
+            logger.info("Forward upsample size to force interpolation output size.")
+            forward_upsample_size = True
+
+        # ensure attention_mask is a bias, and give it a singleton query_tokens dimension
+        # expects mask of shape:
+        #   [batch, key_tokens]
+        # adds singleton query_tokens dimension:
+        #   [batch,                    1, key_tokens]
+        # this helps to broadcast it as a bias over attention scores, which will be in one of the following shapes:
+        #   [batch,  heads, query_tokens, key_tokens] (e.g. torch sdp attn)
+        #   [batch * heads, query_tokens, key_tokens] (e.g. xformers or classic attn)
+        if attention_mask is not None:
+            # assume that mask is expressed as:
+            #   (1 = keep,      0 = discard)
+            # convert mask into a bias that can be added to attention scores:
+            #       (keep = +0,     discard = -10000.0)
+            attention_mask = (1 - attention_mask.to(sample.dtype)) * -10000.0
+            attention_mask = attention_mask.unsqueeze(1)
+
+        # convert encoder_attention_mask to a bias the same way we do for attention_mask
+        if encoder_attention_mask is not None:
+            encoder_attention_mask = (1 - encoder_attention_mask.to(sample.dtype)) * -10000.0
+            encoder_attention_mask = encoder_attention_mask.unsqueeze(1)
+
+        # 0. center input if necessary
+        if self.config.center_input_sample:
+            sample = 2 * sample - 1.0
+
+        # 1. time
+        timesteps = timestep
+        if not torch.is_tensor(timesteps):
+            # TODO: this requires sync between CPU and GPU. So try to pass timesteps as tensors if you can
+            # This would be a good case for the `match` statement (Python 3.10+)
+            is_mps = sample.device.type == "mps"
+            if isinstance(timestep, float):
+                dtype = torch.float32 if is_mps else torch.float64
+            else:
+                dtype = torch.int32 if is_mps else torch.int64
+            timesteps = torch.tensor([timesteps], dtype=dtype, device=sample.device)
+        elif len(timesteps.shape) == 0:
+            timesteps = timesteps[None].to(sample.device)
+
+        # broadcast to batch dimension in a way that's compatible with ONNX/Core ML
+        timesteps = timesteps.expand(sample.shape[0])
+
+        t_emb = self.time_proj(timesteps)
+
+        # `Timesteps` does not contain any weights and will always return f32 tensors
+        # but time_embedding might actually be running in fp16. so we need to cast here.
+        # there might be better ways to encapsulate this.
+        t_emb = t_emb.to(dtype=sample.dtype)
+
+        emb = self.time_embedding(t_emb, timestep_cond)
+
+        if self.use_frame_stride_condition:
+            if not torch.is_tensor(frame_stride):
+                # TODO: this requires sync between CPU and GPU. So try to pass timesteps as tensors if you can
+                # This would be a good case for the `match` statement (Python 3.10+)
+                is_mps = sample.device.type == "mps"
+                if isinstance(timestep, float):
+                    dtype = torch.float32 if is_mps else torch.float64
+                else:
+                    dtype = torch.int32 if is_mps else torch.int64
+                frame_stride = torch.tensor([frame_stride], dtype=dtype, device=sample.device)
+            elif len(frame_stride.shape) == 0:
+                frame_stride = frame_stride[None].to(sample.device)
+
+            # broadcast to batch dimension in a way that's compatible with ONNX/Core ML
+            frame_stride = frame_stride.expand(sample.shape[0])
+
+            fs_emb = self.time_proj(frame_stride)
+
+            # `Timesteps` does not contain any weights and will always return f32 tensors
+            # but time_embedding might actually be running in fp16. so we need to cast here.
+            # there might be better ways to encapsulate this.
+            fs_emb = fs_emb.to(dtype=sample.dtype)
+
+            fs_emb = self.frame_stride_embedding(fs_emb, timestep_cond)
+            emb = emb + fs_emb
+
+        aug_emb = None
+
+        if self.class_embedding is not None:
+            if class_labels is None:
+                raise ValueError("class_labels should be provided when num_class_embeds > 0")
+
+            if self.config.class_embed_type == "timestep":
+                class_labels = self.time_proj(class_labels)
+
+                # `Timesteps` does not contain any weights and will always return f32 tensors
+                # there might be better ways to encapsulate this.
+                class_labels = class_labels.to(dtype=sample.dtype)
+
+            class_emb = self.class_embedding(class_labels).to(dtype=sample.dtype)
+
+            if self.config.class_embeddings_concat:
+                emb = torch.cat([emb, class_emb], dim=-1)
+            else:
+                emb = emb + class_emb
+
+        if self.config.addition_embed_type == "text":
+            aug_emb = self.add_embedding(encoder_hidden_states)
+        elif self.config.addition_embed_type == "text_image":
+            # Kandinsky 2.1 - style
+            if "image_embeds" not in added_cond_kwargs:
+                raise ValueError(
+                    f"{self.__class__} has the config param `addition_embed_type` set to 'text_image' which requires the keyword argument `image_embeds` to be passed in `added_cond_kwargs`"
+                )
+
+            image_embs = added_cond_kwargs.get("image_embeds")
+            text_embs = added_cond_kwargs.get("text_embeds", encoder_hidden_states)
+            aug_emb = self.add_embedding(text_embs, image_embs)
+        elif self.config.addition_embed_type == "text_time":
+            # SDXL - style
+            if "text_embeds" not in added_cond_kwargs:
+                raise ValueError(
+                    f"{self.__class__} has the config param `addition_embed_type` set to 'text_time' which requires the keyword argument `text_embeds` to be passed in `added_cond_kwargs`"
+                )
+            text_embeds = added_cond_kwargs.get("text_embeds")
+            if "time_ids" not in added_cond_kwargs:
+                raise ValueError(
+                    f"{self.__class__} has the config param `addition_embed_type` set to 'text_time' which requires the keyword argument `time_ids` to be passed in `added_cond_kwargs`"
+                )
+            time_ids = added_cond_kwargs.get("time_ids")
+            time_embeds = self.add_time_proj(time_ids.flatten())
+            time_embeds = time_embeds.reshape((text_embeds.shape[0], -1))
+
+            add_embeds = torch.concat([text_embeds, time_embeds], dim=-1)
+            add_embeds = add_embeds.to(emb.dtype)
+            aug_emb = self.add_embedding(add_embeds)
+        elif self.config.addition_embed_type == "image":
+            # Kandinsky 2.2 - style
+            if "image_embeds" not in added_cond_kwargs:
+                raise ValueError(
+                    f"{self.__class__} has the config param `addition_embed_type` set to 'image' which requires the keyword argument `image_embeds` to be passed in `added_cond_kwargs`"
+                )
+            image_embs = added_cond_kwargs.get("image_embeds")
+            aug_emb = self.add_embedding(image_embs)
+        elif self.config.addition_embed_type == "image_hint":
+            # Kandinsky 2.2 - style
+            if "image_embeds" not in added_cond_kwargs or "hint" not in added_cond_kwargs:
+                raise ValueError(
+                    f"{self.__class__} has the config param `addition_embed_type` set to 'image_hint' which requires the keyword arguments `image_embeds` and `hint` to be passed in `added_cond_kwargs`"
+                )
+            image_embs = added_cond_kwargs.get("image_embeds")
+            hint = added_cond_kwargs.get("hint")
+            aug_emb, hint = self.add_embedding(image_embs, hint)
+            sample = torch.cat([sample, hint], dim=1)
+
+        emb = emb + aug_emb if aug_emb is not None else emb
+
+        if self.time_embed_act is not None:
+            emb = self.time_embed_act(emb)
+
+        if self.encoder_hid_proj is not None and self.config.encoder_hid_dim_type == "text_proj":
+            encoder_hidden_states = self.encoder_hid_proj(encoder_hidden_states)
+        elif self.encoder_hid_proj is not None and self.config.encoder_hid_dim_type == "text_image_proj":
+            # Kadinsky 2.1 - style
+            if "image_embeds" not in added_cond_kwargs:
+                raise ValueError(
+                    f"{self.__class__} has the config param `encoder_hid_dim_type` set to 'text_image_proj' which requires the keyword argument `image_embeds` to be passed in  `added_conditions`"
+                )
+
+            image_embeds = added_cond_kwargs.get("image_embeds")
+            encoder_hidden_states = self.encoder_hid_proj(encoder_hidden_states, image_embeds)
+        elif self.encoder_hid_proj is not None and self.config.encoder_hid_dim_type == "image_proj":
+            # Kandinsky 2.2 - style
+            if "image_embeds" not in added_cond_kwargs:
+                raise ValueError(
+                    f"{self.__class__} has the config param `encoder_hid_dim_type` set to 'image_proj' which requires the keyword argument `image_embeds` to be passed in  `added_conditions`"
+                )
+            image_embeds = added_cond_kwargs.get("image_embeds")
+            encoder_hidden_states = self.encoder_hid_proj(image_embeds)
+        # 2. pre-process
+        sample = self.conv_in(sample)
+
+        # 2.5 GLIGEN position net
+        if cross_attention_kwargs is not None and cross_attention_kwargs.get("gligen", None) is not None:
+            cross_attention_kwargs = cross_attention_kwargs.copy()
+            gligen_args = cross_attention_kwargs.pop("gligen")
+            cross_attention_kwargs["gligen"] = {"objs": self.position_net(**gligen_args)}
+
+        # 3. down
+        lora_scale = cross_attention_kwargs.get("scale", 1.0) if cross_attention_kwargs is not None else 1.0
+
+        is_controlnet = mid_block_additional_residual is not None and down_block_additional_residuals is not None
+        is_adapter = mid_block_additional_residual is None and down_block_additional_residuals is not None
+
+        down_block_res_samples = (sample,)
+        for downsample_block in self.down_blocks:
+            if hasattr(downsample_block, "has_cross_attention") and downsample_block.has_cross_attention:
+                # For t2i-adapter CrossAttnDownBlock2D
+                additional_residuals = {}
+                if is_adapter and len(down_block_additional_residuals) > 0:
+                    additional_residuals["additional_residuals"] = down_block_additional_residuals.pop(0)
+
+                sample, res_samples = downsample_block(
+                    hidden_states=sample,
+                    temb=emb,
+                    encoder_hidden_states=encoder_hidden_states,
+                    attention_mask=attention_mask,
+                    cross_attention_kwargs=cross_attention_kwargs,
+                    encoder_attention_mask=encoder_attention_mask,
+                    first_frame_latents=first_frame_latents,
+                    **additional_residuals,
+                )
+            else:
+                sample, res_samples = downsample_block(hidden_states=sample, temb=emb, scale=lora_scale, first_frame_latents=first_frame_latents,)
+
+                if is_adapter and len(down_block_additional_residuals) > 0:
+                    sample += down_block_additional_residuals.pop(0)
+
+            down_block_res_samples += res_samples
+
+        if is_controlnet:
+            new_down_block_res_samples = ()
+
+            for down_block_res_sample, down_block_additional_residual in zip(
+                down_block_res_samples, down_block_additional_residuals
+            ):
+                down_block_res_sample = down_block_res_sample + down_block_additional_residual
+                new_down_block_res_samples = new_down_block_res_samples + (down_block_res_sample,)
+
+            down_block_res_samples = new_down_block_res_samples
+
+        # 4. mid
+        if self.mid_block is not None:
+            sample = self.mid_block(
+                sample,
+                emb,
+                encoder_hidden_states=encoder_hidden_states,
+                attention_mask=attention_mask,
+                cross_attention_kwargs=cross_attention_kwargs,
+                encoder_attention_mask=encoder_attention_mask,
+                # additional
+                first_frame_latents=first_frame_latents,
+            )
+            # To support T2I-Adapter-XL
+            if (
+                is_adapter
+                and len(down_block_additional_residuals) > 0
+                and sample.shape == down_block_additional_residuals[0].shape
+            ):
+                sample += down_block_additional_residuals.pop(0)
+
+        if is_controlnet:
+            sample = sample + mid_block_additional_residual
+
+        # 5. up
+        for i, upsample_block in enumerate(self.up_blocks):
+            is_final_block = i == len(self.up_blocks) - 1
+
+            res_samples = down_block_res_samples[-len(upsample_block.resnets) :]
+            down_block_res_samples = down_block_res_samples[: -len(upsample_block.resnets)]
+
+            # if we have not reached the final block and need to forward the
+            # upsample size, we do it here
+            if not is_final_block and forward_upsample_size:
+                upsample_size = down_block_res_samples[-1].shape[2:]
+
+            if hasattr(upsample_block, "has_cross_attention") and upsample_block.has_cross_attention:
+                sample = upsample_block(
+                    hidden_states=sample,
+                    temb=emb,
+                    res_hidden_states_tuple=res_samples,
+                    encoder_hidden_states=encoder_hidden_states,
+                    cross_attention_kwargs=cross_attention_kwargs,
+                    upsample_size=upsample_size,
+                    attention_mask=attention_mask,
+                    encoder_attention_mask=encoder_attention_mask,
+                    first_frame_latents=first_frame_latents,
+                )
+            else:
+                sample = upsample_block(
+                    hidden_states=sample,
+                    temb=emb,
+                    res_hidden_states_tuple=res_samples,
+                    upsample_size=upsample_size,
+                    scale=lora_scale,
+                    first_frame_latents=first_frame_latents,
+                )
+
+        # 6. post-process
+        if self.conv_norm_out:
+            sample = self.conv_norm_out(sample)
+            sample = self.conv_act(sample)
+        sample = self.conv_out(sample)
+
+        sample = rearrange(sample, "(b f) c h w -> b c f h w", f=video_length)
+        if self.config.first_frame_condition_mode != "none":
+            sample = sample[:, :, 1:, :, :]
+
+        if not return_dict:
+            return (sample,)
+
+        return UNet2DConditionOutput(sample=sample)
+
+    @classmethod
+    def from_pretrained(cls, pretrained_model_name_or_path: Optional[Union[str, os.PathLike]], **kwargs):
+
+        kwargs.pop("low_cpu_mem_usage", False)
+        kwargs.pop("device_map", None)
+        
+        cache_dir = kwargs.pop("cache_dir", DIFFUSERS_CACHE)
+        ignore_mismatched_sizes = kwargs.pop("ignore_mismatched_sizes", False)
+        force_download = kwargs.pop("force_download", False)
+        from_flax = kwargs.pop("from_flax", False)
+        resume_download = kwargs.pop("resume_download", False)
+        proxies = kwargs.pop("proxies", None)
+        output_loading_info = kwargs.pop("output_loading_info", False)
+        local_files_only = kwargs.pop("local_files_only", HF_HUB_OFFLINE)
+        use_auth_token = kwargs.pop("use_auth_token", None)
+        revision = kwargs.pop("revision", None)
+        torch_dtype = kwargs.pop("torch_dtype", None)
+        subfolder = kwargs.pop("subfolder", None)
+        device_map = None
+        max_memory = kwargs.pop("max_memory", None)
+        offload_folder = kwargs.pop("offload_folder", None)
+        offload_state_dict = kwargs.pop("offload_state_dict", False)
+        low_cpu_mem_usage = False
+        variant = kwargs.pop("variant", None)
+        use_safetensors = kwargs.pop("use_safetensors", None)
+
+        allow_pickle = False
+        if use_safetensors is None:
+            use_safetensors = True
+            allow_pickle = True
+
+        if low_cpu_mem_usage and not is_accelerate_available():
+            low_cpu_mem_usage = False
+            logger.warning(
+                "Cannot initialize model with low cpu memory usage because `accelerate` was not found in the"
+                " environment. Defaulting to `low_cpu_mem_usage=False`. It is strongly recommended to install"
+                " `accelerate` for faster and less memory-intense model loading. You can do so with: \n```\npip"
+                " install accelerate\n```\n."
+            )
+
+        if device_map is not None and not is_accelerate_available():
+            raise NotImplementedError(
+                "Loading and dispatching requires `accelerate`. Please make sure to install accelerate or set"
+                " `device_map=None`. You can install accelerate with `pip install accelerate`."
+            )
+
+        # Check if we can handle device_map and dispatching the weights
+        if device_map is not None and not is_torch_version(">=", "1.9.0"):
+            raise NotImplementedError(
+                "Loading and dispatching requires torch >= 1.9.0. Please either update your PyTorch version or set"
+                " `device_map=None`."
+            )
+
+        if low_cpu_mem_usage is True and not is_torch_version(">=", "1.9.0"):
+            raise NotImplementedError(
+                "Low memory initialization requires torch >= 1.9.0. Please either update your PyTorch version or set"
+                " `low_cpu_mem_usage=False`."
+            )
+
+        if low_cpu_mem_usage is False and device_map is not None:
+            raise ValueError(
+                f"You cannot set `low_cpu_mem_usage` to `False` while using device_map={device_map} for loading and"
+                " dispatching. Please make sure to set `low_cpu_mem_usage=True`."
+            )
+
+        # Load config if we don't provide a configuration
+        config_path = pretrained_model_name_or_path
+
+        user_agent = {
+            "diffusers": __version__,
+            "file_type": "model",
+            "framework": "pytorch",
+        }
+
+        # load config
+        config, unused_kwargs, commit_hash = cls.load_config(
+            config_path,
+            cache_dir=cache_dir,
+            return_unused_kwargs=True,
+            return_commit_hash=True,
+            force_download=force_download,
+            resume_download=resume_download,
+            proxies=proxies,
+            local_files_only=local_files_only,
+            use_auth_token=use_auth_token,
+            revision=revision,
+            subfolder=subfolder,
+            device_map=device_map,
+            max_memory=max_memory,
+            offload_folder=offload_folder,
+            offload_state_dict=offload_state_dict,
+            user_agent=user_agent,
+            **kwargs,
+        )
+
+        # load model
+        model_file = None
+        if from_flax:
+            model_file = _get_model_file(
+                pretrained_model_name_or_path,
+                weights_name=FLAX_WEIGHTS_NAME,
+                cache_dir=cache_dir,
+                force_download=force_download,
+                resume_download=resume_download,
+                proxies=proxies,
+                local_files_only=local_files_only,
+                use_auth_token=use_auth_token,
+                revision=revision,
+                subfolder=subfolder,
+                user_agent=user_agent,
+                commit_hash=commit_hash,
+            )
+            model = cls.from_config(config, **unused_kwargs)
+
+            # Convert the weights
+            from diffusers.models.modeling_pytorch_flax_utils import load_flax_checkpoint_in_pytorch_model
+
+            model = load_flax_checkpoint_in_pytorch_model(model, model_file)
+        else:
+            if use_safetensors:
+                try:
+                    model_file = _get_model_file(
+                        pretrained_model_name_or_path,
+                        weights_name=_add_variant(SAFETENSORS_WEIGHTS_NAME, variant),
+                        cache_dir=cache_dir,
+                        force_download=force_download,
+                        resume_download=resume_download,
+                        proxies=proxies,
+                        local_files_only=local_files_only,
+                        use_auth_token=use_auth_token,
+                        revision=revision,
+                        subfolder=subfolder,
+                        user_agent=user_agent,
+                        commit_hash=commit_hash,
+                    )
+                except IOError as e:
+                    if not allow_pickle:
+                        raise e
+                    pass
+            if model_file is None:
+                model_file = _get_model_file(
+                    pretrained_model_name_or_path,
+                    weights_name=_add_variant(WEIGHTS_NAME, variant),
+                    cache_dir=cache_dir,
+                    force_download=force_download,
+                    resume_download=resume_download,
+                    proxies=proxies,
+                    local_files_only=local_files_only,
+                    use_auth_token=use_auth_token,
+                    revision=revision,
+                    subfolder=subfolder,
+                    user_agent=user_agent,
+                    commit_hash=commit_hash,
+                )
+
+            if low_cpu_mem_usage:
+                # Instantiate model with empty weights
+                with accelerate.init_empty_weights():
+                    model = cls.from_config(config, **unused_kwargs)
+
+                # if device_map is None, load the state dict and move the params from meta device to the cpu
+                if device_map is None:
+                    param_device = "cpu"
+                    state_dict = load_state_dict(model_file, variant=variant)
+                    model._convert_deprecated_attention_blocks(state_dict)
+                    # move the params from meta device to cpu
+                    missing_keys = set(model.state_dict().keys()) - set(state_dict.keys())
+                    if len(missing_keys) > 0:
+                        raise ValueError(
+                            f"Cannot load {cls} from {pretrained_model_name_or_path} because the following keys are"
+                            f" missing: \n {', '.join(missing_keys)}. \n Please make sure to pass"
+                            " `low_cpu_mem_usage=False` and `device_map=None` if you want to randomly initialize"
+                            " those weights or else make sure your checkpoint file is correct."
+                        )
+
+                    unexpected_keys = load_model_dict_into_meta(
+                        model,
+                        state_dict,
+                        device=param_device,
+                        dtype=torch_dtype,
+                        model_name_or_path=pretrained_model_name_or_path,
+                    )
+
+                    if cls._keys_to_ignore_on_load_unexpected is not None:
+                        for pat in cls._keys_to_ignore_on_load_unexpected:
+                            unexpected_keys = [k for k in unexpected_keys if re.search(pat, k) is None]
+
+                    if len(unexpected_keys) > 0:
+                        logger.warn(
+                            f"Some weights of the model checkpoint were not used when initializing {cls.__name__}: \n {[', '.join(unexpected_keys)]}"
+                        )
+
+                else:  # else let accelerate handle loading and dispatching.
+                    # Load weights and dispatch according to the device_map
+                    # by default the device_map is None and the weights are loaded on the CPU
+                    try:
+                        accelerate.load_checkpoint_and_dispatch(
+                            model,
+                            model_file,
+                            device_map,
+                            max_memory=max_memory,
+                            offload_folder=offload_folder,
+                            offload_state_dict=offload_state_dict,
+                            dtype=torch_dtype,
+                        )
+                    except AttributeError as e:
+                        # When using accelerate loading, we do not have the ability to load the state
+                        # dict and rename the weight names manually. Additionally, accelerate skips
+                        # torch loading conventions and directly writes into `module.{_buffers, _parameters}`
+                        # (which look like they should be private variables?), so we can't use the standard hooks
+                        # to rename parameters on load. We need to mimic the original weight names so the correct
+                        # attributes are available. After we have loaded the weights, we convert the deprecated
+                        # names to the new non-deprecated names. Then we _greatly encourage_ the user to convert
+                        # the weights so we don't have to do this again.
+
+                        if "'Attention' object has no attribute" in str(e):
+                            logger.warn(
+                                f"Taking `{str(e)}` while using `accelerate.load_checkpoint_and_dispatch` to mean {pretrained_model_name_or_path}"
+                                " was saved with deprecated attention block weight names. We will load it with the deprecated attention block"
+                                " names and convert them on the fly to the new attention block format. Please re-save the model after this conversion,"
+                                " so we don't have to do the on the fly renaming in the future. If the model is from a hub checkpoint,"
+                                " please also re-upload it or open a PR on the original repository."
+                            )
+                            model._temp_convert_self_to_deprecated_attention_blocks()
+                            accelerate.load_checkpoint_and_dispatch(
+                                model,
+                                model_file,
+                                device_map,
+                                max_memory=max_memory,
+                                offload_folder=offload_folder,
+                                offload_state_dict=offload_state_dict,
+                                dtype=torch_dtype,
+                            )
+                            model._undo_temp_convert_self_to_deprecated_attention_blocks()
+                        else:
+                            raise e
+
+                loading_info = {
+                    "missing_keys": [],
+                    "unexpected_keys": [],
+                    "mismatched_keys": [],
+                    "error_msgs": [],
+                }
+            else:
+                model = cls.from_config(config, **unused_kwargs)
+
+                state_dict = load_state_dict(model_file, variant=variant)
+                model._convert_deprecated_attention_blocks(state_dict)
+
+                model, missing_keys, unexpected_keys, mismatched_keys, error_msgs = cls._load_pretrained_model(
+                    model,
+                    state_dict,
+                    model_file,
+                    pretrained_model_name_or_path,
+                    ignore_mismatched_sizes=ignore_mismatched_sizes,
+                )
+
+                loading_info = {
+                    "missing_keys": missing_keys,
+                    "unexpected_keys": unexpected_keys,
+                    "mismatched_keys": mismatched_keys,
+                    "error_msgs": error_msgs,
+                }
+
+        if torch_dtype is not None and not isinstance(torch_dtype, torch.dtype):
+            raise ValueError(
+                f"{torch_dtype} needs to be of type `torch.dtype`, e.g. `torch.float16`, but is {type(torch_dtype)}."
+            )
+        elif torch_dtype is not None:
+            model = model.to(torch_dtype)
+
+        model.register_to_config(_name_or_path=pretrained_model_name_or_path)
+
+        m, u = loading_info["missing_keys"], loading_info["unexpected_keys"]
+        logger.info(f"### missing keys: {len(m)}; unexpected keys: {len(u)};")
+        # print(f"### missing keys:\n{m}\n### unexpected keys:\n{u}\n")
+
+        spatial_params = [p.numel() if "conv3ds" not in n and "tempo_attns" not in n else 0 for n, p in model.named_parameters()]
+        tconv_params = [p.numel() if "conv3ds." in n else 0 for n, p in model.named_parameters()]
+        tattn_params = [p.numel() if "tempo_attns." in n else 0 for n, p in model.named_parameters()]
+        tffconv_params = [p.numel() if "first_frame_conv." in n else 0 for n, p in model.named_parameters()]
+        logger.info(f"### First Frame Convolution Layer Parameters: {sum(tffconv_params) / 1e6} M")
+        logger.info(f"### Spatial UNet Parameters: {sum(spatial_params) / 1e6} M")
+        logger.info(f"### Temporal Convolution Module Parameters: {sum(tconv_params) / 1e6} M")
+        logger.info(f"### Temporal Attention Module Parameters: {sum(tattn_params) / 1e6} M")
+
+        # Set model in evaluation mode to deactivate DropOut modules by default
+        model.eval()
+        if output_loading_info:
+            return model, loading_info
+
+        return model
+
+if __name__ == "__main__":
+    # test
+    from diffusers import AutoencoderKL, DDIMScheduler
+    from transformers import CLIPTextModel, CLIPTokenizer
+    from consisti2v.pipelines.pipeline_animation import AnimationPipeline
+    from consisti2v.pipelines.pipeline_conditional_animation import ConditionalAnimationPipeline
+    from consisti2v.utils.util import save_videos_grid
+
+    pretrained_model_path = "models/StableDiffusion/stable-diffusion-v1-5"
+    prompt = "apply eye makeup"
+    first_frame_path = "/ML-A100/home/weiming/datasets/UCF/frames/v_ApplyEyeMakeup_g01_c01_frame_90.jpg"
+
+    tokenizer    = CLIPTokenizer.from_pretrained(pretrained_model_path, subfolder="tokenizer", use_safetensors=True)
+    text_encoder = CLIPTextModel.from_pretrained(pretrained_model_path, subfolder="text_encoder")
+    vae          = AutoencoderKL.from_pretrained(pretrained_model_path, subfolder="vae", use_safetensors=True)            
+    unet         = VideoLDMUNet3DConditionModel.from_pretrained(
+        pretrained_model_path,
+        subfolder="unet",
+        use_safetensors=True
+    )
+
+    noise_scheduler_kwargs = {
+        "num_train_timesteps": 1000,
+        "beta_start":          0.00085,
+        "beta_end":            0.012,
+        "beta_schedule":       "linear",
+        "steps_offset":        1,
+        "clip_sample":         False,
+    }
+    noise_scheduler = DDIMScheduler(**noise_scheduler_kwargs)
+    # latent = torch.randn(1, 4, 8, 64, 64).to("cuda")
+    # text_embedding = torch.randn(1, 77, 768).to("cuda")
+    # timestep = torch.randint(0, 1000, (1,)).to("cuda").squeeze(0)
+    # output = unet(latent, timestep, text_embedding)
+
+    pipeline = ConditionalAnimationPipeline(
+        unet=unet, vae=vae, tokenizer=tokenizer, text_encoder=text_encoder, scheduler=noise_scheduler,
+    ).to("cuda")
+    sample = pipeline(
+        prompt,
+        num_inference_steps = 25,
+        guidance_scale      = 8.,
+        video_length        = 8,
+        height              = 256,
+        width               =  256,
+        first_frame_paths   = first_frame_path,
+    ).videos
+    print(sample.shape)
+    save_videos_grid(sample, f"samples/videoldm.gif")

consisti2v/models/videoldm_unet_blocks.py

@@ -0,0 +1,1159 @@
+from typing import Optional, Dict, Tuple, Any
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from einops import rearrange, repeat
+from einops.layers.torch import Rearrange
+from diffusers.utils import logging
+from diffusers.models.unet_2d_blocks import (
+    DownBlock2D,
+    UpBlock2D
+)
+from diffusers.models.resnet import (
+    ResnetBlock2D,
+    Downsample2D,
+    Upsample2D,
+)
+from diffusers.models.transformer_2d import Transformer2DModelOutput
+from diffusers.models.dual_transformer_2d import DualTransformer2DModel
+from diffusers.models.activations import get_activation
+from diffusers.utils import logging, is_torch_version
+from diffusers.utils.import_utils import is_xformers_available
+from .videoldm_transformer_blocks import Transformer2DConditionModel
+
+logger = logging.get_logger(__name__)
+
+if is_xformers_available():
+    import xformers
+    import xformers.ops
+else:
+    xformers = None
+
+
+def get_down_block(
+    down_block_type,
+    num_layers,
+    in_channels,
+    out_channels,
+    temb_channels,
+    add_downsample,
+    resnet_eps,
+    resnet_act_fn,
+    transformer_layers_per_block=1,
+    num_attention_heads=None,
+    resnet_groups=None,
+    cross_attention_dim=None,
+    downsample_padding=None,
+    dual_cross_attention=False,
+    use_linear_projection=False,
+    only_cross_attention=False,
+    upcast_attention=False,
+    resnet_time_scale_shift="default",
+    attention_type="default",
+    resnet_skip_time_act=False,
+    resnet_out_scale_factor=1.0,
+    cross_attention_norm=None,
+    attention_head_dim=None,
+    downsample_type=None,
+    dropout=0.0,
+    # additional
+    use_temporal=True,
+    augment_temporal_attention=False,
+    n_frames=8,
+    n_temp_heads=8,
+    first_frame_condition_mode="none",
+    latent_channels=4,
+    rotary_emb=False,
+):
+    # If attn head dim is not defined, we default it to the number of heads
+    if attention_head_dim is None:
+        logger.warn(
+            f"It is recommended to provide `attention_head_dim` when calling `get_down_block`. Defaulting `attention_head_dim` to {num_attention_heads}."
+        )
+        attention_head_dim = num_attention_heads
+        
+    down_block_type = down_block_type[7:] if down_block_type.startswith("UNetRes") else down_block_type
+    if down_block_type == "DownBlock2D":
+        return VideoLDMDownBlock(
+            num_layers=num_layers,
+            in_channels=in_channels,
+            out_channels=out_channels,
+            temb_channels=temb_channels,
+            dropout=dropout,
+            add_downsample=add_downsample,
+            resnet_eps=resnet_eps,
+            resnet_act_fn=resnet_act_fn,
+            resnet_groups=resnet_groups,
+            downsample_padding=downsample_padding,
+            resnet_time_scale_shift=resnet_time_scale_shift,
+            # additional
+            use_temporal=use_temporal,
+            n_frames=n_frames,
+            first_frame_condition_mode=first_frame_condition_mode,
+            latent_channels=latent_channels
+        )
+    elif down_block_type == "CrossAttnDownBlock2D":
+        return VideoLDMCrossAttnDownBlock(
+            num_layers=num_layers,
+            transformer_layers_per_block=transformer_layers_per_block,
+            in_channels=in_channels,
+            out_channels=out_channels,
+            temb_channels=temb_channels,
+            dropout=dropout,
+            add_downsample=add_downsample,
+            resnet_eps=resnet_eps,
+            resnet_act_fn=resnet_act_fn,
+            resnet_groups=resnet_groups,
+            downsample_padding=downsample_padding,
+            cross_attention_dim=cross_attention_dim,
+            num_attention_heads=num_attention_heads,
+            dual_cross_attention=dual_cross_attention,
+            use_linear_projection=use_linear_projection,
+            only_cross_attention=only_cross_attention,
+            upcast_attention=upcast_attention,
+            resnet_time_scale_shift=resnet_time_scale_shift,
+            attention_type=attention_type,
+            # additional
+            use_temporal=use_temporal,
+            augment_temporal_attention=augment_temporal_attention,
+            n_frames=n_frames,
+            n_temp_heads=n_temp_heads,
+            first_frame_condition_mode=first_frame_condition_mode,
+            latent_channels=latent_channels,
+            rotary_emb=rotary_emb,
+        )
+
+    raise ValueError(f'{down_block_type} does not exist.')
+
+
+def get_up_block(
+    up_block_type,
+    num_layers,
+    in_channels,
+    out_channels,
+    prev_output_channel,
+    temb_channels,
+    add_upsample,
+    resnet_eps,
+    resnet_act_fn,
+    transformer_layers_per_block=1,
+    num_attention_heads=None,
+    resnet_groups=None,
+    cross_attention_dim=None,
+    dual_cross_attention=False,
+    use_linear_projection=False,
+    only_cross_attention=False,
+    upcast_attention=False,
+    resnet_time_scale_shift="default",
+    attention_type="default",
+    resnet_skip_time_act=False,
+    resnet_out_scale_factor=1.0,
+    cross_attention_norm=None,
+    attention_head_dim=None,
+    upsample_type=None,
+    dropout=0.0,
+    # additional
+    use_temporal=True,
+    augment_temporal_attention=False,
+    n_frames=8,
+    n_temp_heads=8,
+    first_frame_condition_mode="none",
+    latent_channels=4,
+    rotary_emb=None,
+):
+    if attention_head_dim is None:
+        logger.warn(
+            f"It is recommended to provide `attention_head_dim` when calling `get_up_block`. Defaulting `attention_head_dim` to {num_attention_heads}."
+        )
+        attention_head_dim = num_attention_heads
+
+    up_block_type = up_block_type[7:] if up_block_type.startswith("UNetRes") else up_block_type
+    if up_block_type == "UpBlock2D":
+        return VideoLDMUpBlock(
+            num_layers=num_layers,
+            in_channels=in_channels,
+            out_channels=out_channels,
+            prev_output_channel=prev_output_channel,
+            temb_channels=temb_channels,
+            dropout=dropout,
+            add_upsample=add_upsample,
+            resnet_eps=resnet_eps,
+            resnet_act_fn=resnet_act_fn,
+            resnet_groups=resnet_groups,
+            resnet_time_scale_shift=resnet_time_scale_shift,
+            # additional
+            use_temporal=use_temporal,
+            n_frames=n_frames,
+            first_frame_condition_mode=first_frame_condition_mode,
+            latent_channels=latent_channels
+        )
+    elif up_block_type == 'CrossAttnUpBlock2D':
+        return VideoLDMCrossAttnUpBlock(
+            num_layers=num_layers,
+            transformer_layers_per_block=transformer_layers_per_block,
+            in_channels=in_channels,
+            out_channels=out_channels,
+            prev_output_channel=prev_output_channel,
+            temb_channels=temb_channels,
+            dropout=dropout,
+            add_upsample=add_upsample,
+            resnet_eps=resnet_eps,
+            resnet_act_fn=resnet_act_fn,
+            resnet_groups=resnet_groups,
+            cross_attention_dim=cross_attention_dim,
+            num_attention_heads=num_attention_heads,
+            dual_cross_attention=dual_cross_attention,
+            use_linear_projection=use_linear_projection,
+            only_cross_attention=only_cross_attention,
+            upcast_attention=upcast_attention,
+            resnet_time_scale_shift=resnet_time_scale_shift,
+            attention_type=attention_type,
+            # additional
+            use_temporal=use_temporal,
+            augment_temporal_attention=augment_temporal_attention,
+            n_frames=n_frames,
+            n_temp_heads=n_temp_heads,
+            first_frame_condition_mode=first_frame_condition_mode,
+            latent_channels=latent_channels,
+            rotary_emb=rotary_emb,
+        )
+
+    raise ValueError(f'{up_block_type} does not exist.')
+
+
+class TemporalResnetBlock(nn.Module):
+    def __init__(
+        self,
+        *,
+        in_channels,
+        out_channels=None,
+        dropout=0.0,
+        temb_channels=512,
+        groups=32,
+        groups_out=None,
+        pre_norm=True,
+        eps=1e-6,
+        non_linearity="swish",
+        time_embedding_norm="default",
+        output_scale_factor=1.0,
+        # additional
+        n_frames=8,
+    ):
+        super().__init__()
+        self.pre_norm = pre_norm
+        self.pre_norm = True
+        self.in_channels = in_channels
+        out_channels = in_channels if out_channels is None else out_channels
+        self.out_channels = out_channels
+        self.time_embedding_norm = time_embedding_norm
+        self.output_scale_factor = output_scale_factor
+
+        if groups_out is None:
+            groups_out = groups
+
+        self.norm1 = torch.nn.GroupNorm(num_groups=groups, num_channels=in_channels, eps=eps, affine=True)
+
+        self.conv1 = Conv3DLayer(in_channels, out_channels, n_frames=n_frames)
+
+        if temb_channels is not None:
+            if self.time_embedding_norm == "default":
+                time_emb_proj_out_channels = out_channels
+            elif self.time_embedding_norm == "scale_shift":
+                time_emb_proj_out_channels = out_channels * 2
+            else:
+                raise ValueError(f"unknown time_embedding_norm : {self.time_embedding_norm} ")
+
+            self.time_emb_proj = torch.nn.Linear(temb_channels, time_emb_proj_out_channels)
+        else:
+            self.time_emb_proj = None
+
+        self.norm2 = torch.nn.GroupNorm(num_groups=groups_out, num_channels=out_channels, eps=eps, affine=True)
+
+        self.dropout = torch.nn.Dropout(dropout)
+        self.conv2 = Conv3DLayer(out_channels, out_channels, n_frames=n_frames)
+
+        self.nonlinearity = get_activation(non_linearity)
+
+        self.alpha = nn.Parameter(torch.ones(1))
+
+    def forward(self, input_tensor, temb=None):
+        hidden_states = input_tensor
+
+        hidden_states = self.norm1(hidden_states)
+        hidden_states = self.nonlinearity(hidden_states)
+
+        hidden_states = self.conv1(hidden_states)
+
+        if temb is not None:
+            temb = self.time_emb_proj(self.nonlinearity(temb))[:, :, None, None, None]
+
+        if temb is not None and self.time_embedding_norm == "default":
+            hidden_states = hidden_states + temb
+
+        hidden_states = self.norm2(hidden_states)
+
+        if temb is not None and self.time_embedding_norm == "scale_shift":
+            scale, shift = torch.chunk(temb, 2, dim=1)
+            hidden_states = hidden_states * (1 + scale) + shift
+
+        hidden_states = self.nonlinearity(hidden_states)
+
+        hidden_states = self.dropout(hidden_states)
+        hidden_states = self.conv2(hidden_states)
+
+        output_tensor = (input_tensor + hidden_states) / self.output_scale_factor
+
+        # weighted sum between spatial and temporal features
+        with torch.no_grad():
+            self.alpha.clamp_(0, 1)
+
+        output_tensor = self.alpha * input_tensor + (1 - self.alpha) * output_tensor
+
+        return output_tensor
+
+
+class Conv3DLayer(nn.Conv3d):
+    def __init__(self, in_dim, out_dim, n_frames):
+        k, p = (3, 1, 1), (1, 0, 0)
+        super().__init__(in_channels=in_dim, out_channels=out_dim, kernel_size=k, stride=1, padding=p)
+
+        self.to_3d = Rearrange('(b t) c h w -> b c t h w', t=n_frames)
+        self.to_2d = Rearrange('b c t h w -> (b t) c h w')
+
+    def forward(self, x):
+        h = self.to_3d(x)
+        h = super().forward(h)
+        out = self.to_2d(h)
+        return out
+
+
+class IdentityLayer(nn.Identity):
+    def __init__(self, return_trans2d_output, *args, **kwargs):
+        super().__init__()
+        self.return_trans2d_output = return_trans2d_output
+
+    def forward(self, x, *args, **kwargs):
+        if self.return_trans2d_output:
+            return Transformer2DModelOutput(sample=x)
+        else:
+            return x
+
+
+class VideoLDMCrossAttnDownBlock(nn.Module):
+    def __init__(
+        self,
+        in_channels: int,
+        out_channels: int,
+        temb_channels: int,
+        dropout: float = 0.0,
+        num_layers: int = 1,
+        transformer_layers_per_block: int = 1,
+        resnet_eps: float = 1e-6,
+        resnet_time_scale_shift: str = "default",
+        resnet_act_fn: str = "swish",
+        resnet_groups: int = 32,
+        resnet_pre_norm: bool = True,
+        num_attention_heads=1,
+        cross_attention_dim=1280,
+        output_scale_factor=1.0,
+        downsample_padding=1,
+        add_downsample=True,
+        dual_cross_attention=False,
+        use_linear_projection=False,
+        only_cross_attention=False,
+        upcast_attention=False,
+        attention_type="default",
+        # additional
+        use_temporal=True,
+        augment_temporal_attention=False,
+        n_frames=8,
+        n_temp_heads=8,
+        first_frame_condition_mode="none",
+        latent_channels=4,
+        rotary_emb=False,
+    ):
+        super().__init__()
+
+        self.use_temporal = use_temporal
+
+        self.n_frames = n_frames
+        self.first_frame_condition_mode = first_frame_condition_mode
+        if self.first_frame_condition_mode == "conv2d":
+            self.first_frame_conv = nn.Conv2d(latent_channels, in_channels, kernel_size=1)
+
+        resnets = []
+        attentions = []
+
+        self.n_frames = n_frames
+        self.n_temp_heads = n_temp_heads
+
+        self.has_cross_attention = True
+        self.num_attention_heads = num_attention_heads
+
+        for i in range(num_layers):
+            in_channels = in_channels if i == 0 else out_channels
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=output_scale_factor,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+            if not dual_cross_attention:
+                attentions.append(
+                    Transformer2DConditionModel(
+                        num_attention_heads,
+                        out_channels // num_attention_heads,
+                        in_channels=out_channels,
+                        num_layers=transformer_layers_per_block,
+                        cross_attention_dim=cross_attention_dim,
+                        norm_num_groups=resnet_groups,
+                        use_linear_projection=use_linear_projection,
+                        only_cross_attention=only_cross_attention,
+                        upcast_attention=upcast_attention,
+                        attention_type=attention_type,
+                        # additional
+                        n_frames=n_frames,
+                    )
+                )
+            else:
+                attentions.append(
+                    DualTransformer2DModel(
+                        num_attention_heads,
+                        out_channels // num_attention_heads,
+                        in_channels=out_channels,
+                        num_layers=1,
+                        cross_attention_dim=cross_attention_dim,
+                        norm_num_groups=resnet_groups,
+                    )
+                )
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+        if add_downsample:
+            self.downsamplers = nn.ModuleList(
+                [
+                    Downsample2D(
+                        out_channels, use_conv=True, out_channels=out_channels, padding=downsample_padding, name="op"
+                    )
+                ]
+            )
+        else:
+            self.downsamplers = None
+
+        self.gradient_checkpointing = False
+
+        # >>> Temporal Layers >>>
+        conv3ds = []
+        tempo_attns = []
+
+        for i in range(num_layers):
+            if self.use_temporal:
+                conv3ds.append(
+                    TemporalResnetBlock(
+                        in_channels=out_channels,
+                        out_channels=out_channels,
+                        n_frames=n_frames,
+                    )
+                )
+
+                tempo_attns.append(
+                    Transformer2DConditionModel(
+                        n_temp_heads,
+                        out_channels // n_temp_heads,
+                        in_channels=out_channels,
+                        num_layers=transformer_layers_per_block,
+                        cross_attention_dim=cross_attention_dim,
+                        norm_num_groups=resnet_groups,
+                        use_linear_projection=use_linear_projection,
+                        only_cross_attention=only_cross_attention,
+                        upcast_attention=upcast_attention,
+                        attention_type=attention_type,
+                        # additional
+                        n_frames=n_frames,
+                        is_temporal=True,
+                        augment_temporal_attention=augment_temporal_attention,
+                        rotary_emb=rotary_emb
+                    )
+                )
+            else:
+                conv3ds.append(IdentityLayer(return_trans2d_output=False))
+                tempo_attns.append(IdentityLayer(return_trans2d_output=True))
+
+        self.conv3ds = nn.ModuleList(conv3ds)
+        self.tempo_attns = nn.ModuleList(tempo_attns)
+        # <<< Temporal Layers <<<
+
+    def forward(
+        self,
+        hidden_states: torch.FloatTensor,
+        temb: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.FloatTensor] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        # additional
+        first_frame_latents=None,
+    ):
+        condition_on_first_frame = (self.first_frame_condition_mode != "none" and self.first_frame_condition_mode != "input_only")
+        # input shape: hidden_states = (b f) c h w, first_frame_latents = b c 1 h w
+        if self.first_frame_condition_mode == "conv2d":
+            hidden_states = rearrange(hidden_states, '(b t) c h w -> b c t h w', t=self.n_frames)
+            hidden_height = hidden_states.shape[3]
+            first_frame_height = first_frame_latents.shape[3]
+            downsample_ratio = hidden_height / first_frame_height
+            first_frame_latents = F.interpolate(first_frame_latents.squeeze(2), scale_factor=downsample_ratio, mode="nearest")
+            first_frame_latents = self.first_frame_conv(first_frame_latents).unsqueeze(2)
+            hidden_states[:, :, 0:1, :, :] = first_frame_latents
+            hidden_states = rearrange(hidden_states, 'b c t h w -> (b t) c h w', t=self.n_frames)
+
+        output_states = ()
+
+        for resnet, conv3d, attn, tempo_attn in zip(self.resnets, self.conv3ds, self.attentions, self.tempo_attns):
+
+            hidden_states = resnet(hidden_states, temb)
+            hidden_states = conv3d(hidden_states)
+            hidden_states = attn(
+                hidden_states,
+                encoder_hidden_states=encoder_hidden_states,
+                cross_attention_kwargs=cross_attention_kwargs,
+                condition_on_first_frame=condition_on_first_frame,
+            ).sample
+            hidden_states = tempo_attn(
+                hidden_states,
+                encoder_hidden_states=encoder_hidden_states,
+                cross_attention_kwargs=cross_attention_kwargs,
+                condition_on_first_frame=False,
+            ).sample
+
+            output_states += (hidden_states,)
+
+        if self.downsamplers is not None:
+            for downsampler in self.downsamplers:
+                hidden_states = downsampler(hidden_states)
+
+            output_states += (hidden_states,)
+
+        return hidden_states, output_states
+
+
+class VideoLDMCrossAttnUpBlock(nn.Module):
+    def __init__(
+        self,
+        in_channels: int,
+        out_channels: int,
+        prev_output_channel: int,
+        temb_channels: int,
+        dropout: float = 0.0,
+        num_layers: int = 1,
+        transformer_layers_per_block: int = 1,
+        resnet_eps: float = 1e-6,
+        resnet_time_scale_shift: str = "default",
+        resnet_act_fn: str = "swish",
+        resnet_groups: int = 32,
+        resnet_pre_norm: bool = True,
+        num_attention_heads=1,
+        cross_attention_dim=1280,
+        output_scale_factor=1.0,
+        add_upsample=True,
+        dual_cross_attention=False,
+        use_linear_projection=False,
+        only_cross_attention=False,
+        upcast_attention=False,
+        attention_type="default",
+        # additional
+        use_temporal=True,
+        augment_temporal_attention=False,
+        n_frames=8,
+        n_temp_heads=8,
+        first_frame_condition_mode="none",
+        latent_channels=4,
+        rotary_emb=False,
+    ):
+        super().__init__()
+
+        self.use_temporal = use_temporal
+
+        self.n_frames = n_frames
+        self.first_frame_condition_mode = first_frame_condition_mode
+        if self.first_frame_condition_mode == "conv2d":
+            self.first_frame_conv = nn.Conv2d(latent_channels, prev_output_channel, kernel_size=1)
+
+        resnets = []
+        attentions = []
+
+        self.n_frames = n_frames
+        self.n_temp_heads = n_temp_heads
+
+        self.has_cross_attention = True
+        self.num_attention_heads = num_attention_heads
+
+        for i in range(num_layers):
+            res_skip_channels = in_channels if (i == num_layers - 1) else out_channels
+            resnet_in_channels = prev_output_channel if i == 0 else out_channels
+
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=resnet_in_channels + res_skip_channels,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=output_scale_factor,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+            if not dual_cross_attention:
+                attentions.append(
+                    Transformer2DConditionModel(
+                        num_attention_heads,
+                        out_channels // num_attention_heads,
+                        in_channels=out_channels,
+                        num_layers=transformer_layers_per_block,
+                        cross_attention_dim=cross_attention_dim,
+                        norm_num_groups=resnet_groups,
+                        use_linear_projection=use_linear_projection,
+                        only_cross_attention=only_cross_attention,
+                        upcast_attention=upcast_attention,
+                        attention_type=attention_type,
+                        # additional
+                        n_frames=n_frames,
+                    )
+                )
+            else:
+                attentions.append(
+                    DualTransformer2DModel(
+                        num_attention_heads,
+                        out_channels // num_attention_heads,
+                        in_channels=out_channels,
+                        num_layers=1,
+                        cross_attention_dim=cross_attention_dim,
+                        norm_num_groups=resnet_groups,
+                    )
+                )
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+        if add_upsample:
+            self.upsamplers = nn.ModuleList([Upsample2D(out_channels, use_conv=True, out_channels=out_channels)])
+        else:
+            self.upsamplers = None
+
+        self.gradient_checkpointing = False
+
+        # >>> Temporal Layers >>>
+        conv3ds = []
+        tempo_attns = []
+
+        for i in range(num_layers):
+            if self.use_temporal:
+                conv3ds.append(
+                    TemporalResnetBlock(
+                        in_channels=out_channels,
+                        out_channels=out_channels,
+                        n_frames=n_frames,
+                    )
+                )
+
+                tempo_attns.append(
+                    Transformer2DConditionModel(
+                        n_temp_heads,
+                        out_channels // n_temp_heads,
+                        in_channels=out_channels,
+                        num_layers=transformer_layers_per_block,
+                        cross_attention_dim=cross_attention_dim,
+                        norm_num_groups=resnet_groups,
+                        use_linear_projection=use_linear_projection,
+                        only_cross_attention=only_cross_attention,
+                        upcast_attention=upcast_attention,
+                        attention_type=attention_type,
+                        # additional
+                        n_frames=n_frames,
+                        augment_temporal_attention=augment_temporal_attention,
+                        is_temporal=True,
+                        rotary_emb=rotary_emb,
+                    )
+                )
+            else:
+                conv3ds.append(IdentityLayer(return_trans2d_output=False))
+                tempo_attns.append(IdentityLayer(return_trans2d_output=True))
+
+        self.conv3ds = nn.ModuleList(conv3ds)
+        self.tempo_attns = nn.ModuleList(tempo_attns)
+        # <<< Temporal Layers <<<
+
+    def forward(
+        self,
+        hidden_states: torch.FloatTensor,
+        res_hidden_states_tuple: Tuple[torch.FloatTensor, ...],
+        temb: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.FloatTensor] = None,
+        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
+        upsample_size: Optional[int] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        # additional
+        first_frame_latents=None,
+    ):
+        condition_on_first_frame = (self.first_frame_condition_mode != "none" and self.first_frame_condition_mode != "input_only")
+        # input shape: hidden_states = (b f) c h w, first_frame_latents = b c 1 h w
+        if self.first_frame_condition_mode == "conv2d":
+            hidden_states = rearrange(hidden_states, '(b t) c h w -> b c t h w', t=self.n_frames)
+            hidden_height = hidden_states.shape[3]
+            first_frame_height = first_frame_latents.shape[3]
+            downsample_ratio = hidden_height / first_frame_height
+            first_frame_latents = F.interpolate(first_frame_latents.squeeze(2), scale_factor=downsample_ratio, mode="nearest")
+            first_frame_latents = self.first_frame_conv(first_frame_latents).unsqueeze(2)
+            hidden_states[:, :, 0:1, :, :] = first_frame_latents
+            hidden_states = rearrange(hidden_states, 'b c t h w -> (b t) c h w', t=self.n_frames)
+
+        for resnet, conv3d, attn, tempo_attn in zip(self.resnets, self.conv3ds, self.attentions, self.tempo_attns):
+
+            res_hidden_states = res_hidden_states_tuple[-1]
+            res_hidden_states_tuple = res_hidden_states_tuple[:-1]
+            hidden_states = torch.cat([hidden_states, res_hidden_states], dim=1)
+
+            hidden_states = resnet(hidden_states, temb)
+            hidden_states = conv3d(hidden_states)
+            hidden_states = attn(
+                hidden_states,
+                encoder_hidden_states=encoder_hidden_states,
+                cross_attention_kwargs=cross_attention_kwargs,
+                condition_on_first_frame=condition_on_first_frame,
+            ).sample
+            hidden_states = tempo_attn(
+                hidden_states,
+                encoder_hidden_states=encoder_hidden_states,
+                cross_attention_kwargs=cross_attention_kwargs,
+                condition_on_first_frame=False,
+            ).sample
+
+        if self.upsamplers is not None:
+            for upsampler in self.upsamplers:
+                hidden_states = upsampler(hidden_states, upsample_size)
+        return hidden_states
+    
+
+class VideoLDMUNetMidBlock2DCrossAttn(nn.Module):
+    def __init__(
+        self,
+        in_channels: int,
+        temb_channels: int,
+        dropout: float = 0.0,
+        num_layers: int = 1,
+        transformer_layers_per_block: int = 1,
+        resnet_eps: float = 1e-6,
+        resnet_time_scale_shift: str = "default",
+        resnet_act_fn: str = "swish",
+        resnet_groups: int = 32,
+        resnet_pre_norm: bool = True,
+        num_attention_heads=1,
+        output_scale_factor=1.0,
+        cross_attention_dim=1280,
+        dual_cross_attention=False,
+        use_linear_projection=False,
+        upcast_attention=False,
+        attention_type="default",
+        # additional
+        use_temporal=True,
+        n_frames: int = 8,
+        first_frame_condition_mode="none",
+        latent_channels=4,
+    ):
+        super().__init__()
+
+        self.use_temporal = use_temporal
+
+        self.n_frames = n_frames
+        self.first_frame_condition_mode = first_frame_condition_mode
+        if self.first_frame_condition_mode == "conv2d":
+            self.first_frame_conv = nn.Conv2d(latent_channels, in_channels, kernel_size=1)
+
+        self.has_cross_attention = True
+        self.num_attention_heads = num_attention_heads
+        resnet_groups = resnet_groups if resnet_groups is not None else min(in_channels // 4, 32)
+
+        # there is always at least one resnet
+        resnets = [
+            ResnetBlock2D(
+                in_channels=in_channels,
+                out_channels=in_channels,
+                temb_channels=temb_channels,
+                eps=resnet_eps,
+                groups=resnet_groups,
+                dropout=dropout,
+                time_embedding_norm=resnet_time_scale_shift,
+                non_linearity=resnet_act_fn,
+                output_scale_factor=output_scale_factor,
+                pre_norm=resnet_pre_norm,
+            )
+        ]
+        if self.use_temporal:
+            conv3ds = [
+                TemporalResnetBlock(
+                    in_channels=in_channels,
+                    out_channels=in_channels,
+                    n_frames=n_frames,
+                )
+            ]
+        else:
+            conv3ds = [IdentityLayer(return_trans2d_output=False)]
+
+        attentions = []
+
+        for _ in range(num_layers):
+            if not dual_cross_attention:
+                attentions.append(
+                    Transformer2DConditionModel(
+                        num_attention_heads,
+                        in_channels // num_attention_heads,
+                        in_channels=in_channels,
+                        num_layers=transformer_layers_per_block,
+                        cross_attention_dim=cross_attention_dim,
+                        norm_num_groups=resnet_groups,
+                        use_linear_projection=use_linear_projection,
+                        upcast_attention=upcast_attention,
+                        attention_type=attention_type,
+                        # additional
+                        n_frames=n_frames,
+                    )
+                )
+            else:
+                attentions.append(
+                    DualTransformer2DModel(
+                        num_attention_heads,
+                        in_channels // num_attention_heads,
+                        in_channels=in_channels,
+                        num_layers=1,
+                        cross_attention_dim=cross_attention_dim,
+                        norm_num_groups=resnet_groups,
+                    )
+                )
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels,
+                    out_channels=in_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=output_scale_factor,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+            if self.use_temporal:
+                conv3ds.append(
+                    TemporalResnetBlock(
+                        in_channels=in_channels,
+                        out_channels=in_channels,
+                        n_frames=n_frames,
+                    )
+                )
+            else:
+                conv3ds.append(IdentityLayer(return_trans2d_output=False))
+
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+        self.conv3ds = nn.ModuleList(conv3ds)
+
+        self.gradient_checkpointing = False
+
+    def forward(
+        self,
+        hidden_states: torch.FloatTensor,
+        temb: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.FloatTensor] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        # additional
+        first_frame_latents=None,
+    ) -> torch.FloatTensor:
+        condition_on_first_frame = (self.first_frame_condition_mode != "none" and self.first_frame_condition_mode != "input_only")
+        # input shape: hidden_states = (b f) c h w, first_frame_latents = b c 1 h w
+        if self.first_frame_condition_mode == "conv2d":
+            hidden_states = rearrange(hidden_states, '(b t) c h w -> b c t h w', t=self.n_frames)
+            hidden_height = hidden_states.shape[3]
+            first_frame_height = first_frame_latents.shape[3]
+            downsample_ratio = hidden_height / first_frame_height
+            first_frame_latents = F.interpolate(first_frame_latents.squeeze(2), scale_factor=downsample_ratio, mode="nearest")
+            first_frame_latents = self.first_frame_conv(first_frame_latents).unsqueeze(2)
+            hidden_states[:, :, 0:1, :, :] = first_frame_latents
+            hidden_states = rearrange(hidden_states, 'b c t h w -> (b t) c h w', t=self.n_frames)
+
+        lora_scale = cross_attention_kwargs.get("scale", 1.0) if cross_attention_kwargs is not None else 1.0
+        hidden_states = self.resnets[0](hidden_states, temb, scale=lora_scale)
+        hidden_states = self.conv3ds[0](hidden_states)
+        for attn, resnet, conv3d in zip(self.attentions, self.resnets[1:], self.conv3ds[1:]):
+            if self.training and self.gradient_checkpointing:
+
+                def create_custom_forward(module, return_dict=None):
+                    def custom_forward(*inputs):
+                        if return_dict is not None:
+                            return module(*inputs, return_dict=return_dict)
+                        else:
+                            return module(*inputs)
+
+                    return custom_forward
+
+                ckpt_kwargs: Dict[str, Any] = {"use_reentrant": False} if is_torch_version(">=", "1.11.0") else {}
+                hidden_states = attn(
+                    hidden_states,
+                    encoder_hidden_states=encoder_hidden_states,
+                    cross_attention_kwargs=cross_attention_kwargs,
+                    attention_mask=attention_mask,
+                    encoder_attention_mask=encoder_attention_mask,
+                    return_dict=False,
+                    # additional
+                    condition_on_first_frame=condition_on_first_frame,
+                )[0]
+                hidden_states = torch.utils.checkpoint.checkpoint(
+                    create_custom_forward(resnet),
+                    hidden_states,
+                    temb,
+                    **ckpt_kwargs,
+                )
+                hidden_states = conv3d(hidden_states)
+            else:
+                hidden_states = attn(
+                    hidden_states,
+                    encoder_hidden_states=encoder_hidden_states,
+                    cross_attention_kwargs=cross_attention_kwargs,
+                    attention_mask=attention_mask,
+                    encoder_attention_mask=encoder_attention_mask,
+                    return_dict=False,
+                    # additional
+                    condition_on_first_frame=condition_on_first_frame,
+                )[0]
+                hidden_states = resnet(hidden_states, temb, scale=lora_scale)
+                hidden_states = conv3d(hidden_states)
+
+        return hidden_states
+
+
+class VideoLDMDownBlock(DownBlock2D):
+    def __init__(
+        self,
+        in_channels: int,
+        out_channels: int,
+        temb_channels: int,
+        dropout: float = 0.0,
+        num_layers: int = 1,
+        resnet_eps: float = 1e-6,
+        resnet_time_scale_shift: str = "default",
+        resnet_act_fn: str = "swish",
+        resnet_groups: int = 32,
+        resnet_pre_norm: bool = True,
+        output_scale_factor=1.0,
+        add_downsample=True,
+        downsample_padding=1,
+        # additional
+        use_temporal=True,
+        n_frames: int = 8,
+        first_frame_condition_mode="none",
+        latent_channels=4,
+    ):
+        super().__init__(
+            in_channels,
+            out_channels,
+            temb_channels,
+            dropout,
+            num_layers,
+            resnet_eps,
+            resnet_time_scale_shift,
+            resnet_act_fn,
+            resnet_groups,
+            resnet_pre_norm,
+            output_scale_factor,
+            add_downsample,
+            downsample_padding,)
+
+        self.use_temporal = use_temporal
+
+        self.n_frames = n_frames
+        self.first_frame_condition_mode = first_frame_condition_mode
+        if self.first_frame_condition_mode == "conv2d":
+            self.first_frame_conv = nn.Conv2d(latent_channels, in_channels, kernel_size=1)
+
+        # >>> Temporal Layers >>>
+        conv3ds = []
+        for i in range(num_layers):
+            if self.use_temporal:
+                conv3ds.append(
+                    TemporalResnetBlock(
+                        in_channels=out_channels,
+                        out_channels=out_channels,
+                        n_frames=n_frames,
+                    )
+                )
+            else:
+                conv3ds.append(IdentityLayer(return_trans2d_output=False))
+        self.conv3ds = nn.ModuleList(conv3ds)
+        # <<< Temporal Layers <<<
+
+    def forward(self, hidden_states, temb=None, scale: float = 1, first_frame_latents=None):
+        # input shape: hidden_states = (b f) c h w, first_frame_latents = b c 1 h w
+        if self.first_frame_condition_mode == "conv2d":
+            hidden_states = rearrange(hidden_states, '(b t) c h w -> b c t h w', t=self.n_frames)
+            hidden_height = hidden_states.shape[3]
+            first_frame_height = first_frame_latents.shape[3]
+            downsample_ratio = hidden_height / first_frame_height
+            first_frame_latents = F.interpolate(first_frame_latents.squeeze(2), scale_factor=downsample_ratio, mode="nearest")
+            first_frame_latents = self.first_frame_conv(first_frame_latents).unsqueeze(2)
+            hidden_states[:, :, 0:1, :, :] = first_frame_latents
+            hidden_states = rearrange(hidden_states, 'b c t h w -> (b t) c h w', t=self.n_frames)
+
+        output_states = ()
+
+        for resnet, conv3d in zip(self.resnets, self.conv3ds):
+            if self.training and self.gradient_checkpointing:
+
+                def create_custom_forward(module):
+                    def custom_forward(*inputs):
+                        return module(*inputs)
+
+                    return custom_forward
+
+                if is_torch_version(">=", "1.11.0"):
+                    hidden_states = torch.utils.checkpoint.checkpoint(
+                        create_custom_forward(resnet), hidden_states, temb, use_reentrant=False
+                    )
+                else:
+                    hidden_states = torch.utils.checkpoint.checkpoint(
+                        create_custom_forward(resnet), hidden_states, temb
+                    )
+            else:
+                hidden_states = resnet(hidden_states, temb, scale=scale)
+
+            hidden_states = conv3d(hidden_states)
+
+            output_states = output_states + (hidden_states,)
+
+        if self.downsamplers is not None:
+            for downsampler in self.downsamplers:
+                hidden_states = downsampler(hidden_states, scale=scale)
+
+            output_states = output_states + (hidden_states,)
+
+        return hidden_states, output_states
+    
+
+class VideoLDMUpBlock(UpBlock2D):
+    def __init__(
+        self,
+        in_channels: int,
+        prev_output_channel: int,
+        out_channels: int,
+        temb_channels: int,
+        dropout: float = 0.0,
+        num_layers: int = 1,
+        resnet_eps: float = 1e-6,
+        resnet_time_scale_shift: str = "default",
+        resnet_act_fn: str = "swish",
+        resnet_groups: int = 32,
+        resnet_pre_norm: bool = True,
+        output_scale_factor=1.0,
+        add_upsample=True,
+        # additional
+        use_temporal=True,
+        n_frames: int = 8,
+        first_frame_condition_mode="none",
+        latent_channels=4,
+    ):
+        super().__init__(
+            in_channels,
+            prev_output_channel,
+            out_channels,
+            temb_channels,
+            dropout,
+            num_layers,
+            resnet_eps,
+            resnet_time_scale_shift,
+            resnet_act_fn,
+            resnet_groups,
+            resnet_pre_norm,
+            output_scale_factor,
+            add_upsample,
+        )
+
+        self.use_temporal = use_temporal
+
+        self.n_frames = n_frames
+        self.first_frame_condition_mode = first_frame_condition_mode
+        if self.first_frame_condition_mode == "conv2d":
+            self.first_frame_conv = nn.Conv2d(latent_channels, prev_output_channel, kernel_size=1)
+
+        # >>> Temporal Layers >>>
+        conv3ds = []
+        for i in range(num_layers):
+            if self.use_temporal:
+                conv3ds.append(
+                    TemporalResnetBlock(
+                        in_channels=out_channels,
+                        out_channels=out_channels,
+                        n_frames=n_frames,
+                    )
+                )
+            else:
+                conv3ds.append(IdentityLayer(return_trans2d_output=False))
+
+        self.conv3ds = nn.ModuleList(conv3ds)
+        # <<< Temporal Layers <<<
+
+    def forward(self, hidden_states, res_hidden_states_tuple, temb=None, upsample_size=None, scale: float = 1, first_frame_latents=None):
+        # input shape: hidden_states = (b f) c h w, first_frame_latents = b c 1 h w
+        if self.first_frame_condition_mode == "conv2d":
+            hidden_states = rearrange(hidden_states, '(b t) c h w -> b c t h w', t=self.n_frames)
+            hidden_height = hidden_states.shape[3]
+            first_frame_height = first_frame_latents.shape[3]
+            downsample_ratio = hidden_height / first_frame_height
+            first_frame_latents = F.interpolate(first_frame_latents.squeeze(2), scale_factor=downsample_ratio, mode="nearest")
+            first_frame_latents = self.first_frame_conv(first_frame_latents).unsqueeze(2)
+            hidden_states[:, :, 0:1, :, :] = first_frame_latents
+            hidden_states = rearrange(hidden_states, 'b c t h w -> (b t) c h w', t=self.n_frames)
+
+        for resnet, conv3d in zip(self.resnets, self.conv3ds):
+            # pop res hidden states
+            res_hidden_states = res_hidden_states_tuple[-1]
+            res_hidden_states_tuple = res_hidden_states_tuple[:-1]
+            hidden_states = torch.cat([hidden_states, res_hidden_states], dim=1)
+
+            if self.training and self.gradient_checkpointing:
+
+                def create_custom_forward(module):
+                    def custom_forward(*inputs):
+                        return module(*inputs)
+
+                    return custom_forward
+
+                if is_torch_version(">=", "1.11.0"):
+                    hidden_states = torch.utils.checkpoint.checkpoint(
+                        create_custom_forward(resnet), hidden_states, temb, use_reentrant=False
+                    )
+                else:
+                    hidden_states = torch.utils.checkpoint.checkpoint(
+                        create_custom_forward(resnet), hidden_states, temb
+                    )
+            else:
+                hidden_states = resnet(hidden_states, temb, scale=scale)
+            
+            hidden_states = conv3d(hidden_states)
+
+        if self.upsamplers is not None:
+            for upsampler in self.upsamplers:
+                hidden_states = upsampler(hidden_states, upsample_size, scale=scale)
+
+        return hidden_states

consisti2v/pipelines/pipeline_autoregress_animation.py

@@ -0,0 +1,615 @@
+# Adapted from https://github.com/showlab/Tune-A-Video/blob/main/tuneavideo/pipelines/pipeline_tuneavideo.py
+
+import inspect
+from typing import Callable, List, Optional, Union
+from dataclasses import dataclass
+
+import math
+import numpy as np
+import torch
+from tqdm import tqdm
+
+from torchvision import transforms as T
+from PIL import Image
+
+from diffusers.utils import is_accelerate_available
+from packaging import version
+from transformers import CLIPTextModel, CLIPTokenizer
+
+from diffusers.configuration_utils import FrozenDict
+from diffusers.models import AutoencoderKL
+from diffusers.pipelines.pipeline_utils import DiffusionPipeline
+from diffusers.schedulers import (
+    DDIMScheduler,
+    DPMSolverMultistepScheduler,
+    EulerAncestralDiscreteScheduler,
+    EulerDiscreteScheduler,
+    LMSDiscreteScheduler,
+    PNDMScheduler,
+)
+from diffusers.utils import deprecate, logging, BaseOutput
+
+from einops import rearrange, repeat
+
+from ..models.unet import UNet3DConditionModel
+from ..utils.frameinit_utils import freq_mix_3d, get_freq_filter
+
+
+logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
+
+# copied from https://github.com/huggingface/diffusers/blob/v0.23.0/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py#L59C1-L70C21
+def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
+    """
+    Rescale `noise_cfg` according to `guidance_rescale`. Based on findings of [Common Diffusion Noise Schedules and
+    Sample Steps are Flawed](https://arxiv.org/pdf/2305.08891.pdf). See Section 3.4
+    """
+    std_text = noise_pred_text.std(dim=list(range(1, noise_pred_text.ndim)), keepdim=True)
+    std_cfg = noise_cfg.std(dim=list(range(1, noise_cfg.ndim)), keepdim=True)
+    # rescale the results from guidance (fixes overexposure)
+    noise_pred_rescaled = noise_cfg * (std_text / std_cfg)
+    # mix with the original results from guidance by factor guidance_rescale to avoid "plain looking" images
+    noise_cfg = guidance_rescale * noise_pred_rescaled + (1 - guidance_rescale) * noise_cfg
+    return noise_cfg
+
+
+@dataclass
+class AnimationPipelineOutput(BaseOutput):
+    videos: Union[torch.Tensor, np.ndarray]
+
+
+class AutoregressiveAnimationPipeline(DiffusionPipeline):
+    _optional_components = []
+
+    def __init__(
+        self,
+        vae: AutoencoderKL,
+        text_encoder: CLIPTextModel,
+        tokenizer: CLIPTokenizer,
+        unet: UNet3DConditionModel,
+        scheduler: Union[
+            DDIMScheduler,
+            PNDMScheduler,
+            LMSDiscreteScheduler,
+            EulerDiscreteScheduler,
+            EulerAncestralDiscreteScheduler,
+            DPMSolverMultistepScheduler,
+        ],
+    ):
+        super().__init__()
+
+        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+            deprecation_message = (
+                f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
+                f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
+                "to update the config accordingly as leaving `steps_offset` might led to incorrect results"
+                " in future versions. If you have downloaded this checkpoint from the Hugging Face Hub,"
+                " it would be very nice if you could open a Pull request for the `scheduler/scheduler_config.json`"
+                " file"
+            )
+            deprecate("steps_offset!=1", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(scheduler.config)
+            new_config["steps_offset"] = 1
+            scheduler._internal_dict = FrozenDict(new_config)
+
+        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+            deprecation_message = (
+                f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
+                " `clip_sample` should be set to False in the configuration file. Please make sure to update the"
+                " config accordingly as not setting `clip_sample` in the config might lead to incorrect results in"
+                " future versions. If you have downloaded this checkpoint from the Hugging Face Hub, it would be very"
+                " nice if you could open a Pull request for the `scheduler/scheduler_config.json` file"
+            )
+            deprecate("clip_sample not set", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(scheduler.config)
+            new_config["clip_sample"] = False
+            scheduler._internal_dict = FrozenDict(new_config)
+
+        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
+            version.parse(unet.config._diffusers_version).base_version
+        ) < version.parse("0.9.0.dev0")
+        is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:
+            deprecation_message = (
+                "The configuration file of the unet has set the default `sample_size` to smaller than"
+                " 64 which seems highly unlikely. If your checkpoint is a fine-tuned version of any of the"
+                " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
+                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
+                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
+                " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
+                " in the config might lead to incorrect results in future versions. If you have downloaded this"
+                " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
+                " the `unet/config.json` file"
+            )
+            deprecate("sample_size<64", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(unet.config)
+            new_config["sample_size"] = 64
+            unet._internal_dict = FrozenDict(new_config)
+
+        self.register_modules(
+            vae=vae,
+            text_encoder=text_encoder,
+            tokenizer=tokenizer,
+            unet=unet,
+            scheduler=scheduler,
+        )
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+
+        self.freq_filter = None
+
+    @torch.no_grad()
+    def init_filter(self, video_length, height, width, filter_params):
+        # initialize frequency filter for noise reinitialization
+        batch_size = 1
+        num_channels_latents = self.unet.config.in_channels
+        filter_shape = [
+            batch_size, 
+            num_channels_latents, 
+            video_length, 
+            height // self.vae_scale_factor, 
+            width // self.vae_scale_factor
+        ]
+        # self.freq_filter = get_freq_filter(filter_shape, device=self._execution_device, params=filter_params)
+        self.freq_filter = get_freq_filter(
+            filter_shape, 
+            device=self._execution_device, 
+            filter_type=filter_params.method,
+            n=filter_params.n if filter_params.method=="butterworth" else None,
+            d_s=filter_params.d_s,
+            d_t=filter_params.d_t
+        )
+
+    def enable_vae_slicing(self):
+        self.vae.enable_slicing()
+
+    def disable_vae_slicing(self):
+        self.vae.disable_slicing()
+
+    def enable_sequential_cpu_offload(self, gpu_id=0):
+        if is_accelerate_available():
+            from accelerate import cpu_offload
+        else:
+            raise ImportError("Please install accelerate via `pip install accelerate`")
+
+        device = torch.device(f"cuda:{gpu_id}")
+
+        for cpu_offloaded_model in [self.unet, self.text_encoder, self.vae]:
+            if cpu_offloaded_model is not None:
+                cpu_offload(cpu_offloaded_model, device)
+
+
+    @property
+    def _execution_device(self):
+        if self.device != torch.device("meta") or not hasattr(self.unet, "_hf_hook"):
+            return self.device
+        for module in self.unet.modules():
+            if (
+                hasattr(module, "_hf_hook")
+                and hasattr(module._hf_hook, "execution_device")
+                and module._hf_hook.execution_device is not None
+            ):
+                return torch.device(module._hf_hook.execution_device)
+        return self.device
+
+    def _encode_prompt(self, prompt, device, num_videos_per_prompt, do_classifier_free_guidance, negative_prompt):
+        batch_size = len(prompt) if isinstance(prompt, list) else 1
+
+        text_inputs = self.tokenizer(
+            prompt,
+            padding="max_length",
+            max_length=self.tokenizer.model_max_length,
+            truncation=True,
+            return_tensors="pt",
+        )
+        text_input_ids = text_inputs.input_ids
+        untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="pt").input_ids
+
+        if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal(text_input_ids, untruncated_ids):
+            removed_text = self.tokenizer.batch_decode(untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1])
+            logger.warning(
+                "The following part of your input was truncated because CLIP can only handle sequences up to"
+                f" {self.tokenizer.model_max_length} tokens: {removed_text}"
+            )
+
+        if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
+            attention_mask = text_inputs.attention_mask.to(device)
+        else:
+            attention_mask = None
+
+        text_embeddings = self.text_encoder(
+            text_input_ids.to(device),
+            attention_mask=attention_mask,
+        )
+        text_embeddings = text_embeddings[0]
+
+        # duplicate text embeddings for each generation per prompt, using mps friendly method
+        bs_embed, seq_len, _ = text_embeddings.shape
+        text_embeddings = text_embeddings.repeat(1, num_videos_per_prompt, 1)
+        text_embeddings = text_embeddings.view(bs_embed * num_videos_per_prompt, seq_len, -1)
+
+        # get unconditional embeddings for classifier free guidance
+        if do_classifier_free_guidance is not None:
+            uncond_tokens: List[str]
+            if negative_prompt is None:
+                uncond_tokens = [""] * batch_size
+            elif type(prompt) is not type(negative_prompt):
+                raise TypeError(
+                    f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !="
+                    f" {type(prompt)}."
+                )
+            elif isinstance(negative_prompt, str):
+                uncond_tokens = [negative_prompt]
+            elif batch_size != len(negative_prompt):
+                raise ValueError(
+                    f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:"
+                    f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"
+                    " the batch size of `prompt`."
+                )
+            else:
+                uncond_tokens = negative_prompt
+
+            max_length = text_input_ids.shape[-1]
+            uncond_input = self.tokenizer(
+                uncond_tokens,
+                padding="max_length",
+                max_length=max_length,
+                truncation=True,
+                return_tensors="pt",
+            )
+
+            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
+                attention_mask = uncond_input.attention_mask.to(device)
+            else:
+                attention_mask = None
+
+            uncond_embeddings = self.text_encoder(
+                uncond_input.input_ids.to(device),
+                attention_mask=attention_mask,
+            )
+            uncond_embeddings = uncond_embeddings[0]
+
+            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
+            seq_len = uncond_embeddings.shape[1]
+            uncond_embeddings = uncond_embeddings.repeat(1, num_videos_per_prompt, 1)
+            uncond_embeddings = uncond_embeddings.view(batch_size * num_videos_per_prompt, seq_len, -1)
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            if do_classifier_free_guidance == "text":
+                text_embeddings = torch.cat([uncond_embeddings, text_embeddings])
+            elif do_classifier_free_guidance == "both":
+                text_embeddings = torch.cat([uncond_embeddings, uncond_embeddings, text_embeddings])
+
+        return text_embeddings
+
+    def decode_latents(self, latents, first_frames=None):
+        video_length = latents.shape[2]
+        latents = 1 / self.vae.config.scaling_factor * latents
+        latents = rearrange(latents, "b c f h w -> (b f) c h w")
+        # video = self.vae.decode(latents).sample
+        video = []
+        for frame_idx in tqdm(range(latents.shape[0]), **self._progress_bar_config):
+            video.append(self.vae.decode(latents[frame_idx:frame_idx+1]).sample)
+        video = torch.cat(video)
+        video = rearrange(video, "(b f) c h w -> b c f h w", f=video_length)
+
+        if first_frames is not None:
+            first_frames = first_frames.unsqueeze(2)
+            video = torch.cat([first_frames, video], dim=2)
+
+        video = (video / 2 + 0.5).clamp(0, 1)
+        # we always cast to float32 as this does not cause significant overhead and is compatible with bfloa16
+        video = video.cpu().float().numpy()
+        return video
+
+    def prepare_extra_step_kwargs(self, generator, eta):
+        # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
+        # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
+        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
+        # and should be between [0, 1]
+
+        accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
+        extra_step_kwargs = {}
+        if accepts_eta:
+            extra_step_kwargs["eta"] = eta
+
+        # check if the scheduler accepts generator
+        accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys())
+        if accepts_generator:
+            extra_step_kwargs["generator"] = generator
+        return extra_step_kwargs
+
+    def check_inputs(self, prompt, height, width, callback_steps, first_frame_paths=None):
+        if not isinstance(prompt, str) and not isinstance(prompt, list):
+            raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")
+
+        if first_frame_paths is not None and (not isinstance(prompt, str) and not isinstance(first_frame_paths, list)):
+            raise ValueError(f"`first_frame_paths` has to be of type `str` or `list` but is {type(first_frame_paths)}")
+
+        if height % 8 != 0 or width % 8 != 0:
+            raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")
+
+        if (callback_steps is None) or (
+            callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)
+        ):
+            raise ValueError(
+                f"`callback_steps` has to be a positive integer but is {callback_steps} of type"
+                f" {type(callback_steps)}."
+            )
+
+    def prepare_latents(self, batch_size, num_channels_latents, video_length, height, width, dtype, device, generator, latents=None, noise_sampling_method="vanilla", noise_alpha=1.0):
+        shape = (batch_size, num_channels_latents, video_length, height // self.vae_scale_factor, width // self.vae_scale_factor)
+        if isinstance(generator, list) and len(generator) != batch_size:
+            raise ValueError(
+                f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
+                f" size of {batch_size}. Make sure the batch size matches the length of the generators."
+            )
+        if latents is None:
+            rand_device = "cpu" if device.type == "mps" else device
+
+            if isinstance(generator, list):
+                # shape = shape
+                shape = (1,) + shape[1:]
+                if noise_sampling_method == "vanilla":
+                    latents = [
+                        torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype)
+                        for i in range(batch_size)
+                    ]
+                elif noise_sampling_method == "pyoco_mixed":
+                    base_shape = (batch_size, num_channels_latents, 1, height // self.vae_scale_factor, width // self.vae_scale_factor)
+                    latents = []
+                    noise_alpha_squared = noise_alpha ** 2
+                    for i in range(batch_size):
+                        base_latent = torch.randn(base_shape, generator=generator[i], device=rand_device, dtype=dtype) * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared))
+                        ind_latent = torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype) * math.sqrt(1 / (1 + noise_alpha_squared))
+                        latents.append(base_latent + ind_latent)
+                elif noise_sampling_method == "pyoco_progressive":
+                    latents = []
+                    noise_alpha_squared = noise_alpha ** 2
+                    for i in range(batch_size):
+                        latent = torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype)
+                        ind_latent = torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype) * math.sqrt(1 / (1 + noise_alpha_squared))
+                        for j in range(1, video_length):
+                            latent[:, :, j, :, :] = latent[:, :, j - 1, :, :] * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared)) + ind_latent[:, :, j, :, :]
+                        latents.append(latent)
+                latents = torch.cat(latents, dim=0).to(device)
+            else:
+                if noise_sampling_method == "vanilla":
+                    latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype).to(device)
+                elif noise_sampling_method == "pyoco_mixed":
+                    noise_alpha_squared = noise_alpha ** 2
+                    base_shape = (batch_size, num_channels_latents, 1, height // self.vae_scale_factor, width // self.vae_scale_factor)
+                    base_latents = torch.randn(base_shape, generator=generator, device=rand_device, dtype=dtype) * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared))
+                    ind_latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype) * math.sqrt(1 / (1 + noise_alpha_squared))
+                    latents = base_latents + ind_latents
+                elif noise_sampling_method == "pyoco_progressive":
+                    noise_alpha_squared = noise_alpha ** 2
+                    latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype)
+                    ind_latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype) * math.sqrt(1 / (1 + noise_alpha_squared))
+                    for j in range(1, video_length):
+                        latents[:, :, j, :, :] = latents[:, :, j - 1, :, :] * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared)) + ind_latents[:, :, j, :, :]
+        else:
+            if latents.shape != shape:
+                raise ValueError(f"Unexpected latents shape, got {latents.shape}, expected {shape}")
+            latents = latents.to(device)
+
+        # scale the initial noise by the standard deviation required by the scheduler
+        latents = latents * self.scheduler.init_noise_sigma
+        return latents
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        prompt: Union[str, List[str]],
+        video_length: Optional[int],
+        height: Optional[int] = None,
+        width: Optional[int] = None,
+        num_inference_steps: int = 50,
+        guidance_scale_txt: float = 7.5,
+        guidance_scale_img: float = 2.0,
+        negative_prompt: Optional[Union[str, List[str]]] = None,
+        num_videos_per_prompt: Optional[int] = 1,
+        eta: float = 0.0,
+        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
+        latents: Optional[torch.FloatTensor] = None,
+        output_type: Optional[str] = "tensor",
+        return_dict: bool = True,
+        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback_steps: Optional[int] = 1,
+        # additional
+        first_frame_paths: Optional[Union[str, List[str]]] = None,
+        first_frames: Optional[torch.FloatTensor] = None,
+        noise_sampling_method: str = "vanilla",
+        noise_alpha: float = 1.0,
+        guidance_rescale: float = 0.0,
+        frame_stride: Optional[int] = None,
+        autoregress_steps: int = 3,
+        use_frameinit: bool = False,
+        frameinit_noise_level: int = 999,
+        **kwargs,
+    ):
+        if first_frame_paths is not None and first_frames is not None:
+            raise ValueError("Only one of `first_frame_paths` and `first_frames` can be passed.")
+        # Default height and width to unet
+        height = height or self.unet.config.sample_size * self.vae_scale_factor
+        width = width or self.unet.config.sample_size * self.vae_scale_factor
+
+        # Check inputs. Raise error if not correct
+        self.check_inputs(prompt, height, width, callback_steps, first_frame_paths)
+
+        # Define call parameters
+        # batch_size = 1 if isinstance(prompt, str) else len(prompt)
+        batch_size = 1
+        if latents is not None:
+            batch_size = latents.shape[0]
+        if isinstance(prompt, list):
+            batch_size = len(prompt)
+            first_frame_input = first_frame_paths if first_frame_paths is not None else first_frames
+            if first_frame_input is not None:
+                assert len(prompt) == len(first_frame_input), "prompt and first_frame_paths should have the same length"
+
+        device = self._execution_device
+        # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
+        # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
+        # corresponds to doing no classifier free guidance.
+        do_classifier_free_guidance = None
+        # two guidance mode: text and text+image
+        if guidance_scale_txt > 1.0:
+            do_classifier_free_guidance = "text"
+        if guidance_scale_img > 1.0:
+            do_classifier_free_guidance = "both"
+
+        # Encode input prompt
+        prompt = prompt if isinstance(prompt, list) else [prompt] * batch_size
+        if negative_prompt is not None:
+            negative_prompt = negative_prompt if isinstance(negative_prompt, list) else [negative_prompt] * batch_size 
+        text_embeddings = self._encode_prompt(
+            prompt, device, num_videos_per_prompt, do_classifier_free_guidance, negative_prompt
+        )
+
+        # Encode input first frame
+        first_frame_latents = None
+        if first_frame_paths is not None:
+            first_frame_paths = first_frame_paths if isinstance(first_frame_paths, list) else [first_frame_paths] * batch_size
+            img_transform = T.Compose([
+                T.ToTensor(),
+                T.Resize(height, antialias=None),
+                T.CenterCrop((height, width)),
+                T.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+            ])
+            first_frames = []
+            for first_frame_path in first_frame_paths:
+                first_frame = Image.open(first_frame_path).convert('RGB')
+                first_frame = img_transform(first_frame).unsqueeze(0)
+                first_frames.append(first_frame)
+            first_frames = torch.cat(first_frames, dim=0)
+        if first_frames is not None:
+            first_frames = first_frames.to(device, dtype=self.vae.dtype)
+            first_frame_latents = self.vae.encode(first_frames).latent_dist
+            first_frame_latents = first_frame_latents.sample()
+            first_frame_latents = first_frame_latents * self.vae.config.scaling_factor # b, c, h, w
+            first_frame_latents = repeat(first_frame_latents, "b c h w -> (b n) c h w", n=num_videos_per_prompt)
+            first_frames = repeat(first_frames, "b c h w -> (b n) c h w", n=num_videos_per_prompt)
+
+        full_video_latent = torch.zeros(batch_size * num_videos_per_prompt, self.unet.config.in_channels, video_length * autoregress_steps - autoregress_steps + 1, height // self.vae_scale_factor, width // self.vae_scale_factor, device=device, dtype=self.vae.dtype)
+
+        start_idx = 0
+        for ar_step in range(autoregress_steps):
+            # Prepare timesteps
+            self.scheduler.set_timesteps(num_inference_steps, device=device)
+            timesteps = self.scheduler.timesteps
+
+            # Prepare latent variables
+            num_channels_latents = self.unet.config.in_channels
+            latents = self.prepare_latents(
+                batch_size * num_videos_per_prompt,
+                num_channels_latents,
+                video_length,
+                height,
+                width,
+                text_embeddings.dtype,
+                device,
+                generator,
+                latents,
+                noise_sampling_method,
+                noise_alpha,
+            )
+            latents_dtype = latents.dtype
+            
+            if use_frameinit:
+                current_diffuse_timestep = frameinit_noise_level # diffuse to noise level
+                diffuse_timesteps = torch.full((batch_size,),int(current_diffuse_timestep))
+                diffuse_timesteps = diffuse_timesteps.long()
+                first_frames_static_vid = repeat(first_frame_latents, "b c h w -> b c t h w", t=video_length)
+                z_T = self.scheduler.add_noise(
+                    original_samples=first_frames_static_vid.to(device), 
+                    noise=latents.to(device), 
+                    timesteps=diffuse_timesteps.to(device)
+                )
+                latents = freq_mix_3d(z_T.to(dtype=torch.float32), latents, LPF=self.freq_filter)
+                latents = latents.to(dtype=latents_dtype)
+            
+            if first_frame_latents is not None:
+                first_frame_noisy_latent = latents[:, :, 0, :, :]
+                latents = latents[:, :, 1:, :, :]
+
+            # Prepare extra step kwargs.
+            extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
+
+            # Denoising loop
+            num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
+            with self.progress_bar(total=num_inference_steps) as progress_bar:
+                for i, t in enumerate(timesteps):
+                    # expand the latents if we are doing classifier free guidance
+                    if do_classifier_free_guidance is None:
+                        latent_model_input = latents
+                    elif do_classifier_free_guidance == "text":
+                        latent_model_input = torch.cat([latents] * 2)
+                    elif do_classifier_free_guidance == "both":
+                        latent_model_input = torch.cat([latents] * 3)
+                    latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
+                    if first_frame_latents is not None:
+                        if do_classifier_free_guidance is None:
+                            first_frame_latents_input = first_frame_latents
+                        elif do_classifier_free_guidance == "text":
+                            first_frame_latents_input = torch.cat([first_frame_latents] * 2)
+                        elif do_classifier_free_guidance == "both":
+                            first_frame_latents_input = torch.cat([first_frame_noisy_latent, first_frame_latents, first_frame_latents])
+
+                        first_frame_latents_input = first_frame_latents_input.unsqueeze(2)
+
+                        # predict the noise residual
+                        noise_pred = self.unet(latent_model_input, t, encoder_hidden_states=text_embeddings, first_frame_latents=first_frame_latents_input, frame_stride=frame_stride).sample.to(dtype=latents_dtype)
+                    else:
+                        noise_pred = self.unet(latent_model_input, t, encoder_hidden_states=text_embeddings).sample.to(dtype=latents_dtype)
+                    # noise_pred = []
+                    # import pdb
+                    # pdb.set_trace()
+                    # for batch_idx in range(latent_model_input.shape[0]):
+                    #     noise_pred_single = self.unet(latent_model_input[batch_idx:batch_idx+1], t, encoder_hidden_states=text_embeddings[batch_idx:batch_idx+1]).sample.to(dtype=latents_dtype)
+                    #     noise_pred.append(noise_pred_single)
+                    # noise_pred = torch.cat(noise_pred)
+
+                    # perform guidance
+                    if do_classifier_free_guidance:
+                        if do_classifier_free_guidance == "text":
+                            noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
+                            noise_pred = noise_pred_uncond + guidance_scale_txt * (noise_pred_text - noise_pred_uncond)
+                        elif do_classifier_free_guidance == "both":
+                            noise_pred_uncond, noise_pred_img, noise_pred_both = noise_pred.chunk(3)
+                            noise_pred = noise_pred_uncond + guidance_scale_img * (noise_pred_img - noise_pred_uncond) + guidance_scale_txt * (noise_pred_both - noise_pred_img)
+                    
+                    if do_classifier_free_guidance and guidance_rescale > 0.0:
+                        # Based on 3.4. in https://arxiv.org/pdf/2305.08891.pdf
+                        # currently only support text guidance
+                        noise_pred = rescale_noise_cfg(noise_pred, noise_pred_text, guidance_rescale=guidance_rescale)
+
+                    # compute the previous noisy sample x_t -> x_t-1
+                    latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs).prev_sample
+
+                    # call the callback, if provided
+                    if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
+                        progress_bar.update()
+                        if callback is not None and i % callback_steps == 0:
+                            callback(i, t, latents)
+
+            # Post-processing
+            
+            latents = torch.cat([first_frame_latents.unsqueeze(2), latents], dim=2)
+            first_frame_latents = latents[:, :, -1, :, :]
+            full_video_latent[:, :, start_idx:start_idx + video_length, :, :] = latents
+
+            latents = None
+            start_idx += (video_length - 1)
+
+        # video = self.decode_latents(latents, first_frames)
+        video = self.decode_latents(full_video_latent)
+
+        # Convert to tensor
+        if output_type == "tensor":
+            video = torch.from_numpy(video)
+
+        if not return_dict:
+            return video
+
+        return AnimationPipelineOutput(videos=video)

consisti2v/pipelines/pipeline_conditional_animation.py

@@ -0,0 +1,695 @@
+# Adapted from https://github.com/showlab/Tune-A-Video/blob/main/tuneavideo/pipelines/pipeline_tuneavideo.py
+
+import inspect
+from typing import Callable, List, Optional, Union
+from dataclasses import dataclass
+
+import math
+import numpy as np
+import torch
+from tqdm import tqdm
+
+from torchvision import transforms as T
+from torchvision.transforms import functional as F
+from PIL import Image
+
+from diffusers.utils import is_accelerate_available
+from packaging import version
+from transformers import CLIPTextModel, CLIPTokenizer
+
+from diffusers.configuration_utils import FrozenDict
+from diffusers.models import AutoencoderKL
+from diffusers.pipelines.pipeline_utils import DiffusionPipeline
+from diffusers.schedulers import (
+    DDIMScheduler,
+    DPMSolverMultistepScheduler,
+    EulerAncestralDiscreteScheduler,
+    EulerDiscreteScheduler,
+    LMSDiscreteScheduler,
+    PNDMScheduler,
+)
+from diffusers.utils import deprecate, logging, BaseOutput
+
+from einops import rearrange, repeat
+
+from ..models.videoldm_unet import VideoLDMUNet3DConditionModel
+
+from ..utils.frameinit_utils import get_freq_filter, freq_mix_3d
+
+
+logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
+
+# copied from https://github.com/huggingface/diffusers/blob/v0.23.0/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py#L59C1-L70C21
+def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
+    """
+    Rescale `noise_cfg` according to `guidance_rescale`. Based on findings of [Common Diffusion Noise Schedules and
+    Sample Steps are Flawed](https://arxiv.org/pdf/2305.08891.pdf). See Section 3.4
+    """
+    std_text = noise_pred_text.std(dim=list(range(1, noise_pred_text.ndim)), keepdim=True)
+    std_cfg = noise_cfg.std(dim=list(range(1, noise_cfg.ndim)), keepdim=True)
+    # rescale the results from guidance (fixes overexposure)
+    noise_pred_rescaled = noise_cfg * (std_text / std_cfg)
+    # mix with the original results from guidance by factor guidance_rescale to avoid "plain looking" images
+    noise_cfg = guidance_rescale * noise_pred_rescaled + (1 - guidance_rescale) * noise_cfg
+    return noise_cfg
+
+def pan_right(image, num_frames=16, crop_width=256):
+    frames = []
+    height, width = image.shape[-2:]
+
+    for i in range(num_frames):
+        # Calculate the start position of the crop
+        start_x = int((width - crop_width) * (i / num_frames))
+        crop = F.crop(image, 0, start_x, height, crop_width)
+        frames.append(crop.unsqueeze(0))
+
+    return torch.cat(frames, dim=0)
+
+
+def pan_left(image, num_frames=16, crop_width=256):
+    frames = []
+    height, width = image.shape[-2:]
+
+    for i in range(num_frames):
+        # Start position moves from right to left
+        start_x = int((width - crop_width) * (1 - (i / num_frames)))
+        crop = F.crop(image, 0, start_x, height, crop_width)
+        frames.append(crop.unsqueeze(0))
+
+    return torch.cat(frames, dim=0)
+
+
+def zoom_in(image, num_frames=16, crop_width=256, ratio=1.5):
+    frames = []
+    height, width = image.shape[-2:]
+    max_crop_size = min(width, height)
+
+    for i in range(num_frames):
+        # Calculate the size of the crop
+        crop_size = max_crop_size - int((max_crop_size - max_crop_size // ratio) * (i / num_frames))
+        start_x = (width - crop_size) // 2
+        start_y = (height - crop_size) // 2
+        crop = F.crop(image, start_y, start_x, crop_size, crop_size)
+        resized_crop = F.resize(crop, (crop_width, crop_width), antialias=None)  # Resize back to original size
+        frames.append(resized_crop.unsqueeze(0))
+
+    return torch.cat(frames, dim=0)
+
+
+def zoom_out(image, num_frames=16, crop_width=256, ratio=1.5):
+    frames = []
+    height, width = image.shape[-2:]
+    min_crop_size = min(width, height) // ratio  # Starting from a quarter of the size
+
+    for i in range(num_frames):
+        # Calculate the size of the crop
+        crop_size = min_crop_size + int((min(width, height) - min_crop_size) * (i / num_frames))
+        start_x = (width - crop_size) // 2
+        start_y = (height - crop_size) // 2
+        crop = F.crop(image, start_y, start_x, crop_size, crop_size)
+        resized_crop = F.resize(crop, (crop_width, crop_width), antialias=None)  # Resize back to original size
+        frames.append(resized_crop.unsqueeze(0))
+
+    return torch.cat(frames, dim=0)
+
+
+@dataclass
+class AnimationPipelineOutput(BaseOutput):
+    videos: Union[torch.Tensor, np.ndarray]
+
+
+class ConditionalAnimationPipeline(DiffusionPipeline):
+    _optional_components = []
+
+    def __init__(
+        self,
+        vae: AutoencoderKL,
+        text_encoder: CLIPTextModel,
+        tokenizer: CLIPTokenizer,
+        unet: VideoLDMUNet3DConditionModel,
+        scheduler: Union[
+            DDIMScheduler,
+            PNDMScheduler,
+            LMSDiscreteScheduler,
+            EulerDiscreteScheduler,
+            EulerAncestralDiscreteScheduler,
+            DPMSolverMultistepScheduler,
+        ],
+    ):
+        super().__init__()
+
+        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+            deprecation_message = (
+                f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
+                f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
+                "to update the config accordingly as leaving `steps_offset` might led to incorrect results"
+                " in future versions. If you have downloaded this checkpoint from the Hugging Face Hub,"
+                " it would be very nice if you could open a Pull request for the `scheduler/scheduler_config.json`"
+                " file"
+            )
+            deprecate("steps_offset!=1", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(scheduler.config)
+            new_config["steps_offset"] = 1
+            scheduler._internal_dict = FrozenDict(new_config)
+
+        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+            deprecation_message = (
+                f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
+                " `clip_sample` should be set to False in the configuration file. Please make sure to update the"
+                " config accordingly as not setting `clip_sample` in the config might lead to incorrect results in"
+                " future versions. If you have downloaded this checkpoint from the Hugging Face Hub, it would be very"
+                " nice if you could open a Pull request for the `scheduler/scheduler_config.json` file"
+            )
+            deprecate("clip_sample not set", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(scheduler.config)
+            new_config["clip_sample"] = False
+            scheduler._internal_dict = FrozenDict(new_config)
+
+        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
+            version.parse(unet.config._diffusers_version).base_version
+        ) < version.parse("0.9.0.dev0")
+        is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:
+            deprecation_message = (
+                "The configuration file of the unet has set the default `sample_size` to smaller than"
+                " 64 which seems highly unlikely. If your checkpoint is a fine-tuned version of any of the"
+                " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
+                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
+                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
+                " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
+                " in the config might lead to incorrect results in future versions. If you have downloaded this"
+                " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
+                " the `unet/config.json` file"
+            )
+            deprecate("sample_size<64", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(unet.config)
+            new_config["sample_size"] = 64
+            unet._internal_dict = FrozenDict(new_config)
+
+        self.register_modules(
+            vae=vae,
+            text_encoder=text_encoder,
+            tokenizer=tokenizer,
+            unet=unet,
+            scheduler=scheduler,
+        )
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+
+        self.freq_filter = None
+
+    @torch.no_grad()
+    def init_filter(self, video_length, height, width, filter_params):
+        # initialize frequency filter for noise reinitialization
+        batch_size = 1
+        num_channels_latents = self.unet.config.in_channels
+        filter_shape = [
+            batch_size, 
+            num_channels_latents, 
+            video_length, 
+            height // self.vae_scale_factor, 
+            width // self.vae_scale_factor
+        ]
+        # self.freq_filter = get_freq_filter(filter_shape, device=self._execution_device, params=filter_params)
+        self.freq_filter = get_freq_filter(
+            filter_shape, 
+            device=self._execution_device, 
+            filter_type=filter_params.method,
+            n=filter_params.n if filter_params.method=="butterworth" else None,
+            d_s=filter_params.d_s,
+            d_t=filter_params.d_t
+        )
+
+    def enable_vae_slicing(self):
+        self.vae.enable_slicing()
+
+    def disable_vae_slicing(self):
+        self.vae.disable_slicing()
+
+    def enable_sequential_cpu_offload(self, gpu_id=0):
+        if is_accelerate_available():
+            from accelerate import cpu_offload
+        else:
+            raise ImportError("Please install accelerate via `pip install accelerate`")
+
+        device = torch.device(f"cuda:{gpu_id}")
+
+        for cpu_offloaded_model in [self.unet, self.text_encoder, self.vae]:
+            if cpu_offloaded_model is not None:
+                cpu_offload(cpu_offloaded_model, device)
+
+
+    @property
+    def _execution_device(self):
+        if self.device != torch.device("meta") or not hasattr(self.unet, "_hf_hook"):
+            return self.device
+        for module in self.unet.modules():
+            if (
+                hasattr(module, "_hf_hook")
+                and hasattr(module._hf_hook, "execution_device")
+                and module._hf_hook.execution_device is not None
+            ):
+                return torch.device(module._hf_hook.execution_device)
+        return self.device
+
+    def _encode_prompt(self, prompt, device, num_videos_per_prompt, do_classifier_free_guidance, negative_prompt):
+        batch_size = len(prompt) if isinstance(prompt, list) else 1
+
+        text_inputs = self.tokenizer(
+            prompt,
+            padding="max_length",
+            max_length=self.tokenizer.model_max_length,
+            truncation=True,
+            return_tensors="pt",
+        )
+        text_input_ids = text_inputs.input_ids
+        untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="pt").input_ids
+
+        if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal(text_input_ids, untruncated_ids):
+            removed_text = self.tokenizer.batch_decode(untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1])
+            logger.warning(
+                "The following part of your input was truncated because CLIP can only handle sequences up to"
+                f" {self.tokenizer.model_max_length} tokens: {removed_text}"
+            )
+
+        if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
+            attention_mask = text_inputs.attention_mask.to(device)
+        else:
+            attention_mask = None
+
+        text_embeddings = self.text_encoder(
+            text_input_ids.to(device),
+            attention_mask=attention_mask,
+        )
+        text_embeddings = text_embeddings[0]
+
+        # duplicate text embeddings for each generation per prompt, using mps friendly method
+        bs_embed, seq_len, _ = text_embeddings.shape
+        text_embeddings = text_embeddings.repeat(1, num_videos_per_prompt, 1)
+        text_embeddings = text_embeddings.view(bs_embed * num_videos_per_prompt, seq_len, -1)
+
+        # get unconditional embeddings for classifier free guidance
+        if do_classifier_free_guidance is not None:
+            uncond_tokens: List[str]
+            if negative_prompt is None:
+                uncond_tokens = [""] * batch_size
+            elif type(prompt) is not type(negative_prompt):
+                raise TypeError(
+                    f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !="
+                    f" {type(prompt)}."
+                )
+            elif isinstance(negative_prompt, str):
+                uncond_tokens = [negative_prompt]
+            elif batch_size != len(negative_prompt):
+                raise ValueError(
+                    f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:"
+                    f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"
+                    " the batch size of `prompt`."
+                )
+            else:
+                uncond_tokens = negative_prompt
+
+            max_length = text_input_ids.shape[-1]
+            uncond_input = self.tokenizer(
+                uncond_tokens,
+                padding="max_length",
+                max_length=max_length,
+                truncation=True,
+                return_tensors="pt",
+            )
+
+            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
+                attention_mask = uncond_input.attention_mask.to(device)
+            else:
+                attention_mask = None
+
+            uncond_embeddings = self.text_encoder(
+                uncond_input.input_ids.to(device),
+                attention_mask=attention_mask,
+            )
+            uncond_embeddings = uncond_embeddings[0]
+
+            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
+            seq_len = uncond_embeddings.shape[1]
+            uncond_embeddings = uncond_embeddings.repeat(1, num_videos_per_prompt, 1)
+            uncond_embeddings = uncond_embeddings.view(batch_size * num_videos_per_prompt, seq_len, -1)
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            if do_classifier_free_guidance == "text":
+                text_embeddings = torch.cat([uncond_embeddings, text_embeddings])
+            elif do_classifier_free_guidance == "both":
+                text_embeddings = torch.cat([uncond_embeddings, uncond_embeddings, text_embeddings])
+
+        return text_embeddings
+
+    def decode_latents(self, latents, first_frames=None):
+        video_length = latents.shape[2]
+        latents = 1 / self.vae.config.scaling_factor * latents
+        latents = rearrange(latents, "b c f h w -> (b f) c h w")
+        # video = self.vae.decode(latents).sample
+        video = []
+        for frame_idx in tqdm(range(latents.shape[0]), **self._progress_bar_config):
+            video.append(self.vae.decode(latents[frame_idx:frame_idx+1]).sample)
+        video = torch.cat(video)
+        video = rearrange(video, "(b f) c h w -> b c f h w", f=video_length)
+
+        if first_frames is not None:
+            first_frames = first_frames.unsqueeze(2)
+            video = torch.cat([first_frames, video], dim=2)
+
+        video = (video / 2 + 0.5).clamp(0, 1)
+        # we always cast to float32 as this does not cause significant overhead and is compatible with bfloa16
+        video = video.cpu().float().numpy()
+        return video
+
+    def prepare_extra_step_kwargs(self, generator, eta):
+        # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
+        # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
+        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
+        # and should be between [0, 1]
+
+        accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
+        extra_step_kwargs = {}
+        if accepts_eta:
+            extra_step_kwargs["eta"] = eta
+
+        # check if the scheduler accepts generator
+        accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys())
+        if accepts_generator:
+            extra_step_kwargs["generator"] = generator
+        return extra_step_kwargs
+
+    def check_inputs(self, prompt, height, width, callback_steps, first_frame_paths=None):
+        if not isinstance(prompt, str) and not isinstance(prompt, list):
+            raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")
+
+        if first_frame_paths is not None and (not isinstance(prompt, str) and not isinstance(first_frame_paths, list)):
+            raise ValueError(f"`first_frame_paths` has to be of type `str` or `list` but is {type(first_frame_paths)}")
+
+        if height % 8 != 0 or width % 8 != 0:
+            raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")
+
+        if (callback_steps is None) or (
+            callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)
+        ):
+            raise ValueError(
+                f"`callback_steps` has to be a positive integer but is {callback_steps} of type"
+                f" {type(callback_steps)}."
+            )
+
+    def prepare_latents(self, batch_size, num_channels_latents, video_length, height, width, dtype, device, generator, latents=None, noise_sampling_method="vanilla", noise_alpha=1.0):
+        shape = (batch_size, num_channels_latents, video_length, height // self.vae_scale_factor, width // self.vae_scale_factor)
+        if isinstance(generator, list) and len(generator) != batch_size:
+            raise ValueError(
+                f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
+                f" size of {batch_size}. Make sure the batch size matches the length of the generators."
+            )
+        if latents is None:
+            rand_device = "cpu" if device.type == "mps" else device
+
+            if isinstance(generator, list):
+                # shape = shape
+                shape = (1,) + shape[1:]
+                if noise_sampling_method == "vanilla":
+                    latents = [
+                        torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype)
+                        for i in range(batch_size)
+                    ]
+                elif noise_sampling_method == "pyoco_mixed":
+                    base_shape = (batch_size, num_channels_latents, 1, height // self.vae_scale_factor, width // self.vae_scale_factor)
+                    latents = []
+                    noise_alpha_squared = noise_alpha ** 2
+                    for i in range(batch_size):
+                        base_latent = torch.randn(base_shape, generator=generator[i], device=rand_device, dtype=dtype) * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared))
+                        ind_latent = torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype) * math.sqrt(1 / (1 + noise_alpha_squared))
+                        latents.append(base_latent + ind_latent)
+                elif noise_sampling_method == "pyoco_progressive":
+                    latents = []
+                    noise_alpha_squared = noise_alpha ** 2
+                    for i in range(batch_size):
+                        latent = torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype)
+                        ind_latent = torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype) * math.sqrt(1 / (1 + noise_alpha_squared))
+                        for j in range(1, video_length):
+                            latent[:, :, j, :, :] = latent[:, :, j - 1, :, :] * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared)) + ind_latent[:, :, j, :, :]
+                        latents.append(latent)
+                latents = torch.cat(latents, dim=0).to(device)
+            else:
+                if noise_sampling_method == "vanilla":
+                    latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype).to(device)
+                elif noise_sampling_method == "pyoco_mixed":
+                    noise_alpha_squared = noise_alpha ** 2
+                    base_shape = (batch_size, num_channels_latents, 1, height // self.vae_scale_factor, width // self.vae_scale_factor)
+                    base_latents = torch.randn(base_shape, generator=generator, device=rand_device, dtype=dtype) * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared))
+                    ind_latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype) * math.sqrt(1 / (1 + noise_alpha_squared))
+                    latents = base_latents + ind_latents
+                elif noise_sampling_method == "pyoco_progressive":
+                    noise_alpha_squared = noise_alpha ** 2
+                    latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype)
+                    ind_latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype) * math.sqrt(1 / (1 + noise_alpha_squared))
+                    for j in range(1, video_length):
+                        latents[:, :, j, :, :] = latents[:, :, j - 1, :, :] * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared)) + ind_latents[:, :, j, :, :]
+        else:
+            if latents.shape != shape:
+                raise ValueError(f"Unexpected latents shape, got {latents.shape}, expected {shape}")
+            latents = latents.to(device)
+
+        # scale the initial noise by the standard deviation required by the scheduler
+        latents = latents * self.scheduler.init_noise_sigma
+        return latents
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        prompt: Union[str, List[str]],
+        video_length: Optional[int],
+        height: Optional[int] = None,
+        width: Optional[int] = None,
+        num_inference_steps: int = 50,
+        guidance_scale_txt: float = 7.5,
+        guidance_scale_img: float = 2.0,
+        negative_prompt: Optional[Union[str, List[str]]] = None,
+        num_videos_per_prompt: Optional[int] = 1,
+        eta: float = 0.0,
+        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
+        latents: Optional[torch.FloatTensor] = None,
+        output_type: Optional[str] = "tensor",
+        return_dict: bool = True,
+        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback_steps: Optional[int] = 1,
+        # additional
+        first_frame_paths: Optional[Union[str, List[str]]] = None,
+        first_frames: Optional[torch.FloatTensor] = None,
+        noise_sampling_method: str = "vanilla",
+        noise_alpha: float = 1.0,
+        guidance_rescale: float = 0.0,
+        frame_stride: Optional[int] = None,
+        use_frameinit: bool = False,
+        frameinit_noise_level: int = 999,
+        camera_motion: str = None,
+        **kwargs,
+    ):
+        if first_frame_paths is not None and first_frames is not None:
+            raise ValueError("Only one of `first_frame_paths` and `first_frames` can be passed.")
+        # Default height and width to unet
+        height = height or self.unet.config.sample_size * self.vae_scale_factor
+        width = width or self.unet.config.sample_size * self.vae_scale_factor
+
+        # Check inputs. Raise error if not correct
+        self.check_inputs(prompt, height, width, callback_steps, first_frame_paths)
+
+        # Define call parameters
+        # batch_size = 1 if isinstance(prompt, str) else len(prompt)
+        batch_size = 1
+        if latents is not None:
+            batch_size = latents.shape[0]
+        if isinstance(prompt, list):
+            batch_size = len(prompt)
+            first_frame_input = first_frame_paths if first_frame_paths is not None else first_frames
+            if first_frame_input is not None:
+                assert len(prompt) == len(first_frame_input), "prompt and first_frame_paths should have the same length"
+
+        device = self._execution_device
+        # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
+        # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
+        # corresponds to doing no classifier free guidance.
+        do_classifier_free_guidance = None
+        # two guidance mode: text and text+image
+        if guidance_scale_txt > 1.0:
+            do_classifier_free_guidance = "text"
+        if guidance_scale_img > 1.0:
+            do_classifier_free_guidance = "both"
+
+        # Encode input prompt
+        prompt = prompt if isinstance(prompt, list) else [prompt] * batch_size
+        if negative_prompt is not None:
+            negative_prompt = negative_prompt if isinstance(negative_prompt, list) else [negative_prompt] * batch_size 
+        text_embeddings = self._encode_prompt(
+            prompt, device, num_videos_per_prompt, do_classifier_free_guidance, negative_prompt
+        )
+
+        # Encode input first frame
+        first_frame_latents = None
+        if first_frame_paths is not None:
+            first_frame_paths = first_frame_paths if isinstance(first_frame_paths, list) else [first_frame_paths] * batch_size
+            if camera_motion is None:
+                img_transform = T.Compose([
+                    T.ToTensor(),
+                    T.Resize(height, antialias=None),
+                    T.CenterCrop((height, width)),
+                    T.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+                ])
+            elif camera_motion == "pan_left" or camera_motion == "pan_right":
+                img_transform = T.Compose([
+                    T.ToTensor(),
+                    T.Resize(height, antialias=None),
+                    T.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+                ])
+            elif camera_motion == "zoom_out" or camera_motion == "zoom_in":
+                img_transform = T.Compose([
+                    T.ToTensor(),
+                    T.Resize(height * 2, antialias=None),
+                    T.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5], inplace=True),
+                ])
+            
+            first_frames = []
+            for first_frame_path in first_frame_paths:
+                first_frame = Image.open(first_frame_path).convert('RGB')
+                first_frame = img_transform(first_frame)
+                if camera_motion is not None:
+                    if camera_motion == "pan_left":
+                        first_frame = pan_left(first_frame, num_frames=video_length, crop_width=width)
+                    elif camera_motion == "pan_right":
+                        first_frame = pan_right(first_frame, num_frames=video_length, crop_width=width)
+                    elif camera_motion == "zoom_in":
+                        first_frame = zoom_in(first_frame, num_frames=video_length, crop_width=width)
+                    elif camera_motion == "zoom_out":
+                        first_frame = zoom_out(first_frame, num_frames=video_length, crop_width=width)
+                    else:
+                        raise NotImplementedError(f"camera_motion: {camera_motion} is not implemented.")
+                first_frames.append(first_frame.unsqueeze(0))
+            first_frames = torch.cat(first_frames, dim=0)
+        if first_frames is not None:
+            first_frames = first_frames.to(device, dtype=self.vae.dtype)
+            if camera_motion is not None:
+                first_frames = rearrange(first_frames, "b f c h w -> (b f) c h w")
+            first_frame_latents = self.vae.encode(first_frames).latent_dist
+            first_frame_latents = first_frame_latents.sample()
+            first_frame_latents = first_frame_latents * self.vae.config.scaling_factor # b, c, h, w
+            first_frame_static_vid = rearrange(first_frame_latents, "(b f) c h w -> b c f h w", f=video_length if camera_motion is not None else 1)
+            first_frame_latents = first_frame_static_vid[:, :, 0, :, :]
+            first_frame_latents = repeat(first_frame_latents, "b c h w -> (b n) c h w", n=num_videos_per_prompt)
+            first_frames = repeat(first_frames, "b c h w -> (b n) c h w", n=num_videos_per_prompt)
+        
+        if use_frameinit and camera_motion is None:
+            first_frame_static_vid = repeat(first_frame_static_vid, "b c 1 h w -> b c t h w", t=video_length)
+
+        # self._progress_bar_config = {}
+        # vid = self.decode_latents(first_frame_static_vid)
+        # vid = torch.from_numpy(vid)
+        # from ..utils.util import save_videos_grid
+        # save_videos_grid(vid, "samples/debug/camera_motion/first_frame_static_vid.mp4", fps=8)
+
+        # Prepare timesteps
+        self.scheduler.set_timesteps(num_inference_steps, device=device)
+        timesteps = self.scheduler.timesteps
+
+        # Prepare latent variables
+        num_channels_latents = self.unet.config.in_channels
+        latents = self.prepare_latents(
+            batch_size * num_videos_per_prompt,
+            num_channels_latents,
+            video_length,
+            height,
+            width,
+            text_embeddings.dtype,
+            device,
+            generator,
+            latents,
+            noise_sampling_method,
+            noise_alpha,
+        )
+        latents_dtype = latents.dtype
+
+        if use_frameinit:
+            current_diffuse_timestep = frameinit_noise_level # diffuse to t noise level
+            diffuse_timesteps = torch.full((batch_size,),int(current_diffuse_timestep))
+            diffuse_timesteps = diffuse_timesteps.long()
+            z_T = self.scheduler.add_noise(
+                original_samples=first_frame_static_vid.to(device), 
+                noise=latents.to(device), 
+                timesteps=diffuse_timesteps.to(device)
+            )
+            latents = freq_mix_3d(z_T.to(dtype=torch.float32), latents.to(dtype=torch.float32), LPF=self.freq_filter)
+            latents = latents.to(dtype=latents_dtype)
+
+        if first_frame_latents is not None:
+            first_frame_noisy_latent = latents[:, :, 0, :, :]
+            latents = latents[:, :, 1:, :, :]
+
+        # Prepare extra step kwargs.
+        extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
+
+        # Denoising loop
+        num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
+        with self.progress_bar(total=num_inference_steps) as progress_bar:
+            for i, t in enumerate(timesteps):
+                # expand the latents if we are doing classifier free guidance
+                if do_classifier_free_guidance is None:
+                    latent_model_input = latents
+                elif do_classifier_free_guidance == "text":
+                    latent_model_input = torch.cat([latents] * 2)
+                elif do_classifier_free_guidance == "both":
+                    latent_model_input = torch.cat([latents] * 3)
+                latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
+                if first_frame_latents is not None:
+                    if do_classifier_free_guidance is None:
+                        first_frame_latents_input = first_frame_latents
+                    elif do_classifier_free_guidance == "text":
+                        first_frame_latents_input = torch.cat([first_frame_latents] * 2)
+                    elif do_classifier_free_guidance == "both":
+                        first_frame_latents_input = torch.cat([first_frame_noisy_latent, first_frame_latents, first_frame_latents])
+
+                    first_frame_latents_input = first_frame_latents_input.unsqueeze(2)
+
+                    # predict the noise residual
+                    noise_pred = self.unet(latent_model_input, t, encoder_hidden_states=text_embeddings, first_frame_latents=first_frame_latents_input, frame_stride=frame_stride).sample.to(dtype=latents_dtype)
+                else:
+                    noise_pred = self.unet(latent_model_input, t, encoder_hidden_states=text_embeddings).sample.to(dtype=latents_dtype)
+
+                # perform guidance
+                if do_classifier_free_guidance:
+                    if do_classifier_free_guidance == "text":
+                        noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
+                        noise_pred = noise_pred_uncond + guidance_scale_txt * (noise_pred_text - noise_pred_uncond)
+                    elif do_classifier_free_guidance == "both":
+                        noise_pred_uncond, noise_pred_img, noise_pred_both = noise_pred.chunk(3)
+                        noise_pred = noise_pred_uncond + guidance_scale_img * (noise_pred_img - noise_pred_uncond) + guidance_scale_txt * (noise_pred_both - noise_pred_img)
+                
+                if do_classifier_free_guidance and guidance_rescale > 0.0:
+                    # Based on 3.4. in https://arxiv.org/pdf/2305.08891.pdf
+                    # currently only support text guidance
+                    noise_pred = rescale_noise_cfg(noise_pred, noise_pred_text, guidance_rescale=guidance_rescale)
+
+                # compute the previous noisy sample x_t -> x_t-1
+                latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs).prev_sample
+
+                # call the callback, if provided
+                if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
+                    progress_bar.update()
+                    if callback is not None and i % callback_steps == 0:
+                        callback(i, t, latents)
+
+        # Post-processing
+        latents = torch.cat([first_frame_latents.unsqueeze(2), latents], dim=2)
+        # video = self.decode_latents(latents, first_frames)
+        video = self.decode_latents(latents)
+
+        # Convert to tensor
+        if output_type == "tensor":
+            video = torch.from_numpy(video)
+
+        if not return_dict:
+            return video
+
+        return AnimationPipelineOutput(videos=video)

consisti2v/utils/frameinit_utils.py

@@ -0,0 +1,142 @@
+# modified from https://github.com/TianxingWu/FreeInit/blob/master/freeinit_utils.py
+import torch
+import torch.fft as fft
+import math
+
+
+def freq_mix_3d(x, noise, LPF):
+    """
+    Noise reinitialization.
+
+    Args:
+        x: diffused latent
+        noise: randomly sampled noise
+        LPF: low pass filter
+    """
+    # FFT
+    x_freq = fft.fftn(x, dim=(-3, -2, -1))
+    x_freq = fft.fftshift(x_freq, dim=(-3, -2, -1))
+    noise_freq = fft.fftn(noise, dim=(-3, -2, -1))
+    noise_freq = fft.fftshift(noise_freq, dim=(-3, -2, -1))
+
+    # frequency mix
+    HPF = 1 - LPF
+    x_freq_low = x_freq * LPF
+    noise_freq_high = noise_freq * HPF
+    x_freq_mixed = x_freq_low + noise_freq_high # mix in freq domain
+
+    # IFFT
+    x_freq_mixed = fft.ifftshift(x_freq_mixed, dim=(-3, -2, -1))
+    x_mixed = fft.ifftn(x_freq_mixed, dim=(-3, -2, -1)).real
+
+    return x_mixed
+
+
+def get_freq_filter(shape, device, filter_type, n, d_s, d_t):
+    """
+    Form the frequency filter for noise reinitialization.
+
+    Args:
+        shape: shape of latent (B, C, T, H, W)
+        filter_type: type of the freq filter
+        n: (only for butterworth) order of the filter, larger n ~ ideal, smaller n ~ gaussian
+        d_s: normalized stop frequency for spatial dimensions (0.0-1.0)
+        d_t: normalized stop frequency for temporal dimension (0.0-1.0)
+    """
+    if filter_type == "gaussian":
+        return gaussian_low_pass_filter(shape=shape, d_s=d_s, d_t=d_t).to(device)
+    elif filter_type == "ideal":
+        return ideal_low_pass_filter(shape=shape, d_s=d_s, d_t=d_t).to(device)
+    elif filter_type == "box":
+        return box_low_pass_filter(shape=shape, d_s=d_s, d_t=d_t).to(device)
+    elif filter_type == "butterworth":
+        return butterworth_low_pass_filter(shape=shape, n=n, d_s=d_s, d_t=d_t).to(device)
+    else:
+        raise NotImplementedError
+
+
+def gaussian_low_pass_filter(shape, d_s=0.25, d_t=0.25):
+    """
+    Compute the gaussian low pass filter mask.
+
+    Args:
+        shape: shape of the filter (volume)
+        d_s: normalized stop frequency for spatial dimensions (0.0-1.0)
+        d_t: normalized stop frequency for temporal dimension (0.0-1.0)
+    """
+    T, H, W = shape[-3], shape[-2], shape[-1]
+    mask = torch.zeros(shape)
+    if d_s==0 or d_t==0:
+        return mask
+    for t in range(T):
+        for h in range(H):
+            for w in range(W):
+                d_square = (((d_s/d_t)*(2*t/T-1))**2 + (2*h/H-1)**2 + (2*w/W-1)**2)
+                mask[..., t,h,w] = math.exp(-1/(2*d_s**2) * d_square)
+    return mask
+
+
+def butterworth_low_pass_filter(shape, n=4, d_s=0.25, d_t=0.25):
+    """
+    Compute the butterworth low pass filter mask.
+
+    Args:
+        shape: shape of the filter (volume)
+        n: order of the filter, larger n ~ ideal, smaller n ~ gaussian
+        d_s: normalized stop frequency for spatial dimensions (0.0-1.0)
+        d_t: normalized stop frequency for temporal dimension (0.0-1.0)
+    """
+    T, H, W = shape[-3], shape[-2], shape[-1]
+    mask = torch.zeros(shape)
+    if d_s==0 or d_t==0:
+        return mask
+    for t in range(T):
+        for h in range(H):
+            for w in range(W):
+                d_square = (((d_s/d_t)*(2*t/T-1))**2 + (2*h/H-1)**2 + (2*w/W-1)**2)
+                mask[..., t,h,w] = 1 / (1 + (d_square / d_s**2)**n)
+    return mask
+
+
+def ideal_low_pass_filter(shape, d_s=0.25, d_t=0.25):
+    """
+    Compute the ideal low pass filter mask.
+
+    Args:
+        shape: shape of the filter (volume)
+        d_s: normalized stop frequency for spatial dimensions (0.0-1.0)
+        d_t: normalized stop frequency for temporal dimension (0.0-1.0)
+    """
+    T, H, W = shape[-3], shape[-2], shape[-1]
+    mask = torch.zeros(shape)
+    if d_s==0 or d_t==0:
+        return mask
+    for t in range(T):
+        for h in range(H):
+            for w in range(W):
+                d_square = (((d_s/d_t)*(2*t/T-1))**2 + (2*h/H-1)**2 + (2*w/W-1)**2)
+                mask[..., t,h,w] =  1 if d_square <= d_s*2 else 0
+    return mask
+
+
+def box_low_pass_filter(shape, d_s=0.25, d_t=0.25):
+    """
+    Compute the ideal low pass filter mask (approximated version).
+
+    Args:
+        shape: shape of the filter (volume)
+        d_s: normalized stop frequency for spatial dimensions (0.0-1.0)
+        d_t: normalized stop frequency for temporal dimension (0.0-1.0)
+    """
+    T, H, W = shape[-3], shape[-2], shape[-1]
+    mask = torch.zeros(shape)
+    if d_s==0 or d_t==0:
+        return mask
+
+    threshold_s = round(int(H // 2) * d_s)
+    threshold_t = round(T // 2 * d_t)
+
+    cframe, crow, ccol = T // 2, H // 2, W //2
+    mask[..., cframe - threshold_t:cframe + threshold_t, crow - threshold_s:crow + threshold_s, ccol - threshold_s:ccol + threshold_s] = 1.0
+
+    return mask

consisti2v/utils/util.py

@@ -0,0 +1,165 @@
+import os
+import imageio
+import numpy as np
+from typing import Union
+
+import torch
+import torchvision
+import torch.distributed as dist
+import wandb
+
+from tqdm import tqdm
+from einops import rearrange
+
+from torchmetrics.image.fid import _compute_fid
+
+
+def zero_rank_print(s):
+    if (not dist.is_initialized()) or (dist.is_initialized() and dist.get_rank() == 0): print("### " + s)
+
+
+def save_videos_grid(videos: torch.Tensor, path: str, rescale=False, n_rows=6, fps=8, wandb=False, global_step=0, format="gif"):
+    videos = rearrange(videos, "b c t h w -> t b c h w")
+    outputs = []
+    for x in videos:
+        x = torchvision.utils.make_grid(x, nrow=n_rows)
+        x = x.transpose(0, 1).transpose(1, 2).squeeze(-1)
+        if rescale:
+            x = (x + 1.0) / 2.0  # -1,1 -> 0,1
+        x = (x * 255).numpy().astype(np.uint8)
+        outputs.append(x)
+
+    if wandb:
+        wandb_video = wandb.Video(outputs, fps=fps)
+        wandb.log({"val_videos": wandb_video}, step=global_step)
+        
+    os.makedirs(os.path.dirname(path), exist_ok=True)
+    if format == "gif":
+        imageio.mimsave(path, outputs, fps=fps)
+    elif format == "mp4":
+        torchvision.io.write_video(path, np.array(outputs), fps=fps, video_codec='h264', options={'crf': '10'})
+
+# DDIM Inversion
+@torch.no_grad()
+def init_prompt(prompt, pipeline):
+    uncond_input = pipeline.tokenizer(
+        [""], padding="max_length", max_length=pipeline.tokenizer.model_max_length,
+        return_tensors="pt"
+    )
+    uncond_embeddings = pipeline.text_encoder(uncond_input.input_ids.to(pipeline.device))[0]
+    text_input = pipeline.tokenizer(
+        [prompt],
+        padding="max_length",
+        max_length=pipeline.tokenizer.model_max_length,
+        truncation=True,
+        return_tensors="pt",
+    )
+    text_embeddings = pipeline.text_encoder(text_input.input_ids.to(pipeline.device))[0]
+    context = torch.cat([uncond_embeddings, text_embeddings])
+
+    return context
+
+
+def next_step(model_output: Union[torch.FloatTensor, np.ndarray], timestep: int,
+              sample: Union[torch.FloatTensor, np.ndarray], ddim_scheduler):
+    timestep, next_timestep = min(
+        timestep - ddim_scheduler.config.num_train_timesteps // ddim_scheduler.num_inference_steps, 999), timestep
+    alpha_prod_t = ddim_scheduler.alphas_cumprod[timestep] if timestep >= 0 else ddim_scheduler.final_alpha_cumprod
+    alpha_prod_t_next = ddim_scheduler.alphas_cumprod[next_timestep]
+    beta_prod_t = 1 - alpha_prod_t
+    next_original_sample = (sample - beta_prod_t ** 0.5 * model_output) / alpha_prod_t ** 0.5
+    next_sample_direction = (1 - alpha_prod_t_next) ** 0.5 * model_output
+    next_sample = alpha_prod_t_next ** 0.5 * next_original_sample + next_sample_direction
+    return next_sample
+
+
+def get_noise_pred_single(latents, t, context, first_frame_latents, frame_stride, unet):
+    noise_pred = unet(latents, t, encoder_hidden_states=context, first_frame_latents=first_frame_latents, frame_stride=frame_stride).sample
+    return noise_pred
+
+
+@torch.no_grad()
+def ddim_loop(pipeline, ddim_scheduler, latent, num_inv_steps, prompt, first_frame_latents, frame_stride):
+    context = init_prompt(prompt, pipeline)
+    uncond_embeddings, cond_embeddings = context.chunk(2)
+    all_latent = [latent]
+    latent = latent.clone().detach()
+    for i in tqdm(range(num_inv_steps)):
+        t = ddim_scheduler.timesteps[len(ddim_scheduler.timesteps) - i - 1]
+        noise_pred = get_noise_pred_single(latent, t, cond_embeddings, first_frame_latents, frame_stride, pipeline.unet)
+        latent = next_step(noise_pred, t, latent, ddim_scheduler)
+        all_latent.append(latent)
+    return all_latent
+
+
+@torch.no_grad()
+def ddim_inversion(pipeline, ddim_scheduler, video_latent, num_inv_steps, prompt="", first_frame_latents=None, frame_stride=3):
+    ddim_latents = ddim_loop(pipeline, ddim_scheduler, video_latent, num_inv_steps, prompt, first_frame_latents, frame_stride)
+    return ddim_latents
+
+
+def compute_fid(real_features, fake_features, num_features, device):
+    orig_dtype = real_features.dtype
+
+    mx_num_feats = (num_features, num_features)
+    real_features_sum = torch.zeros(num_features).double().to(device)
+    real_features_cov_sum = torch.zeros(mx_num_feats).double().to(device)
+    real_features_num_samples = torch.tensor(0).long().to(device)
+
+    fake_features_sum = torch.zeros(num_features).double().to(device)
+    fake_features_cov_sum = torch.zeros(mx_num_feats).double().to(device)
+    fake_features_num_samples = torch.tensor(0).long().to(device)
+
+    real_features = real_features.double()
+    fake_features = fake_features.double()
+
+    real_features_sum += real_features.sum(dim=0)
+    real_features_cov_sum += real_features.t().mm(real_features)
+    real_features_num_samples += real_features.shape[0]
+
+    fake_features_sum += fake_features.sum(dim=0)
+    fake_features_cov_sum += fake_features.t().mm(fake_features)
+    fake_features_num_samples += fake_features.shape[0]
+
+    """Calculate FID score based on accumulated extracted features from the two distributions."""
+    if real_features_num_samples < 2 or fake_features_num_samples < 2:
+        raise RuntimeError("More than one sample is required for both the real and fake distributed to compute FID")
+    mean_real = (real_features_sum / real_features_num_samples).unsqueeze(0)
+    mean_fake = (fake_features_sum / fake_features_num_samples).unsqueeze(0)
+
+    cov_real_num = real_features_cov_sum - real_features_num_samples * mean_real.t().mm(mean_real)
+    cov_real = cov_real_num / (real_features_num_samples - 1)
+    cov_fake_num = fake_features_cov_sum - fake_features_num_samples * mean_fake.t().mm(mean_fake)
+    cov_fake = cov_fake_num / (fake_features_num_samples - 1)
+    return _compute_fid(mean_real.squeeze(0), cov_real, mean_fake.squeeze(0), cov_fake).to(orig_dtype)
+
+
+def compute_inception_score(gen_probs, num_splits=10):
+    num_gen = gen_probs.shape[0]
+    gen_probs = gen_probs.detach().cpu().numpy()
+    scores = []
+    np.random.RandomState(42).shuffle(gen_probs)
+    for i in range(num_splits):
+        part = gen_probs[i * num_gen // num_splits : (i + 1) * num_gen // num_splits]
+        kl = part * (np.log(part) - np.log(np.mean(part, axis=0, keepdims=True)))
+        kl = np.mean(np.sum(kl, axis=1))
+        scores.append(np.exp(kl))
+    return float(np.mean(scores)), float(np.std(scores))
+    # idx = torch.randperm(features.shape[0])
+    # features = features[idx]
+    # # calculate probs and logits
+    # prob = features.softmax(dim=1)
+    # log_prob = features.log_softmax(dim=1)
+
+    # # split into groups
+    # prob = prob.chunk(splits, dim=0)
+    # log_prob = log_prob.chunk(splits, dim=0)
+
+    # # calculate score per split
+    # mean_prob = [p.mean(dim=0, keepdim=True) for p in prob]
+    # kl_ = [p * (log_p - m_p.log()) for p, log_p, m_p in zip(prob, log_prob, mean_prob)]
+    # kl_ = [k.sum(dim=1).mean().exp() for k in kl_]
+    # kl = torch.stack(kl_)
+
+    # return mean and std
+    # return kl.mean(), kl.std()

environment.yaml

@@ -0,0 +1,28 @@
+name: consisti2v
+channels:
+  - pytorch
+  - nvidia
+dependencies:
+  - python=3.10
+  - pytorch=2.1.0
+  - torchvision=0.16.0
+  - torchaudio=2.1.0
+  - pytorch-cuda=11.8
+  - pip
+  - pip:
+    - diffusers==0.21.2
+    - transformers==4.25.1
+    - accelerate==0.23.0
+    - imageio==2.27.0
+    - decord==0.6.0
+    - einops
+    - omegaconf
+    - safetensors
+    - gradio==3.42.0
+    - wandb
+    - moviepy
+    - scikit-learn
+    - av
+    - rotary_embedding_torch
+    - torchmetrics
+    - torch-fidelity

requirements.txt

@@ -0,0 +1,18 @@
+torch==2.1.0
+torchvision==0.16.0
+torchaudio==2.1.0
+diffusers==0.21.2
+transformers==4.25.1
+accelerate==0.23.0
+imageio==2.27.0
+decord==0.6.0
+spaces
+einops
+omegaconf
+safetensors
+moviepy
+scikit-learn
+av
+rotary_embedding_torch
+torchmetrics
+torch-fidelity

scripts/animate.py

@@ -0,0 +1,179 @@
+import argparse
+import datetime
+import random
+import os
+import logging
+from omegaconf import OmegaConf
+
+import torch
+
+import diffusers
+from diffusers import AutoencoderKL, DDIMScheduler
+
+from transformers import CLIPTextModel, CLIPTokenizer
+
+from consisti2v.models.videoldm_unet import VideoLDMUNet3DConditionModel
+from consisti2v.pipelines.pipeline_conditional_animation import ConditionalAnimationPipeline
+from consisti2v.utils.util import save_videos_grid
+from diffusers.utils.import_utils import is_xformers_available
+
+def main(args, config):
+    logging.basicConfig(
+        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+        datefmt="%m/%d/%Y %H:%M:%S",
+        level=logging.INFO,
+    )
+    diffusers.utils.logging.set_verbosity_info()
+
+    time_str = datetime.datetime.now().strftime("%Y-%m-%dT%H-%M-%S")
+    savedir = f"{config.output_dir}/{config.output_name}-{time_str}"
+    os.makedirs(savedir)
+
+    samples = []
+    sample_idx = 0
+
+    ### >>> create validation pipeline >>> ###
+    if config.pipeline_pretrained_path is None:
+        noise_scheduler = DDIMScheduler(**OmegaConf.to_container(config.noise_scheduler_kwargs))
+        tokenizer       = CLIPTokenizer.from_pretrained(config.pretrained_model_path, subfolder="tokenizer", use_safetensors=True)
+        text_encoder    = CLIPTextModel.from_pretrained(config.pretrained_model_path, subfolder="text_encoder")
+        vae             = AutoencoderKL.from_pretrained(config.pretrained_model_path, subfolder="vae", use_safetensors=True)            
+        unet            = VideoLDMUNet3DConditionModel.from_pretrained(
+            config.pretrained_model_path,
+            subfolder="unet",
+            variant=config.unet_additional_kwargs['variant'],
+            temp_pos_embedding=config.unet_additional_kwargs['temp_pos_embedding'],
+            augment_temporal_attention=config.unet_additional_kwargs['augment_temporal_attention'],
+            use_temporal=True,
+            n_frames=config.sampling_kwargs['n_frames'],
+            n_temp_heads=config.unet_additional_kwargs['n_temp_heads'],
+            first_frame_condition_mode=config.unet_additional_kwargs['first_frame_condition_mode'],
+            use_frame_stride_condition=config.unet_additional_kwargs['use_frame_stride_condition'],
+            use_safetensors=True
+        )
+
+        # 1. unet ckpt
+        if config.unet_path is not None:
+            if os.path.isdir(config.unet_path):
+                unet_dict = VideoLDMUNet3DConditionModel.from_pretrained(config.unet_path)
+                m, u = unet.load_state_dict(unet_dict.state_dict(), strict=False)
+                assert len(u) == 0
+                del unet_dict
+            else:
+                checkpoint_dict = torch.load(config.unet_path, map_location="cpu")
+                state_dict = checkpoint_dict["state_dict"] if "state_dict" in checkpoint_dict else checkpoint_dict
+                if config.unet_ckpt_prefix is not None:
+                    state_dict = {k.replace(config.unet_ckpt_prefix, ''): v for k, v in state_dict.items()}
+                m, u = unet.load_state_dict(state_dict, strict=False)
+                assert len(u) == 0
+
+        if is_xformers_available() and int(torch.__version__.split(".")[0]) < 2:
+            unet.enable_xformers_memory_efficient_attention()
+
+        pipeline = ConditionalAnimationPipeline(
+            vae=vae, text_encoder=text_encoder, tokenizer=tokenizer, unet=unet, scheduler=noise_scheduler)
+    
+    else:
+        pipeline = ConditionalAnimationPipeline.from_pretrained(config.pipeline_pretrained_path)
+
+    pipeline.to("cuda")
+
+    # (frameinit) initialize frequency filter for noise reinitialization -------------
+    if config.frameinit_kwargs.enable:
+        pipeline.init_filter(
+            width         = config.sampling_kwargs.width,
+            height        = config.sampling_kwargs.height,
+            video_length  = config.sampling_kwargs.n_frames,
+            filter_params = config.frameinit_kwargs.filter_params,
+        )
+    # -------------------------------------------------------------------------------
+    ### <<< create validation pipeline <<< ###
+
+    if args.prompt is not None:
+        prompts = [args.prompt]
+        n_prompts = [args.n_prompt]
+        first_frame_paths = [args.path_to_first_frame]
+        random_seeds = [int(args.seed)] if args.seed != "random" else "random"
+    else:
+        prompt_config = OmegaConf.load(args.prompt_config)
+        prompts = prompt_config.prompts
+        n_prompts = list(prompt_config.n_prompts) * len(prompts) if len(prompt_config.n_prompts) == 1 else prompt_config.n_prompts
+        first_frame_paths = prompt_config.path_to_first_frames
+        random_seeds = prompt_config.seeds
+    
+    if random_seeds == "random":
+        random_seeds = [random.randint(0, 1e5) for _ in range(len(prompts))]
+    else:
+        random_seeds = [random_seeds] if isinstance(random_seeds, int) else list(random_seeds)
+        random_seeds = random_seeds * len(prompts) if len(random_seeds) == 1 else random_seeds
+    
+    config.prompt_kwargs = OmegaConf.create({"random_seeds": [], "prompts": prompts, "n_prompts": n_prompts, "first_frame_paths": first_frame_paths})
+    for prompt_idx, (prompt, n_prompt, first_frame_path, random_seed) in enumerate(zip(prompts, n_prompts, first_frame_paths, random_seeds)):
+        # manually set random seed for reproduction
+        if random_seed != -1: torch.manual_seed(random_seed)
+        else: torch.seed()
+        config.prompt_kwargs.random_seeds.append(torch.initial_seed())
+        
+        print(f"current seed: {torch.initial_seed()}")
+        print(f"sampling {prompt} ...")
+        sample = pipeline(
+            prompt,
+            negative_prompt       = n_prompt,
+            first_frame_paths     = first_frame_path,
+            num_inference_steps   = config.sampling_kwargs.steps,
+            guidance_scale_txt    = config.sampling_kwargs.guidance_scale_txt,
+            guidance_scale_img    = config.sampling_kwargs.guidance_scale_img,
+            width                 = config.sampling_kwargs.width,
+            height                = config.sampling_kwargs.height,
+            video_length          = config.sampling_kwargs.n_frames,
+            noise_sampling_method = config.unet_additional_kwargs['noise_sampling_method'],
+            noise_alpha           = float(config.unet_additional_kwargs['noise_alpha']),
+            eta                   = config.sampling_kwargs.ddim_eta,
+            frame_stride          = config.sampling_kwargs.frame_stride,
+            guidance_rescale      = config.sampling_kwargs.guidance_rescale,
+            num_videos_per_prompt = config.sampling_kwargs.num_videos_per_prompt,
+            use_frameinit         = config.frameinit_kwargs.enable,
+            frameinit_noise_level = config.frameinit_kwargs.noise_level,
+            camera_motion         = config.frameinit_kwargs.camera_motion,
+        ).videos
+        samples.append(sample)
+
+        prompt = "-".join((prompt.replace("/", "").split(" ")[:10])).replace(":", "")
+        if sample.shape[0] > 1:
+            for cnt, samp in enumerate(sample):
+                save_videos_grid(samp.unsqueeze(0), f"{savedir}/sample/{sample_idx}-{cnt + 1}-{prompt}.{args.format}", format=args.format)
+        else:
+            save_videos_grid(sample, f"{savedir}/sample/{sample_idx}-{prompt}.{args.format}", format=args.format)
+        print(f"save to {savedir}/sample/{prompt}.{args.format}")
+        
+        sample_idx += 1
+
+    samples = torch.concat(samples)
+    save_videos_grid(samples, f"{savedir}/sample.{args.format}", n_rows=4, format=args.format)
+
+    OmegaConf.save(config, f"{savedir}/config.yaml")
+
+    if args.save_model:
+        pipeline.save_pretrained(f"{savedir}/model")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--inference_config", type=str, default="configs/inference/inference.yaml")
+    parser.add_argument("--prompt", "-p", type=str, default=None)
+    parser.add_argument("--n_prompt", "-n", type=str, default="")
+    parser.add_argument("--seed", type=str, default="random")
+    parser.add_argument("--path_to_first_frame", "-f", type=str, default=None)
+    parser.add_argument("--prompt_config", type=str, default="configs/prompts/default.yaml")
+    parser.add_argument("--format", type=str, default="mp4", choices=["gif", "mp4"])
+    parser.add_argument("--save_model", action="store_true")
+    parser.add_argument("optional_args", nargs='*', default=[])
+    args = parser.parse_args()
+
+    config = OmegaConf.load(args.inference_config)
+
+    if args.optional_args:
+        modified_config = OmegaConf.from_dotlist(args.optional_args)
+        config = OmegaConf.merge(config, modified_config)
+
+    main(args, config)

scripts/animate_autoregress.py

@@ -0,0 +1,185 @@
+import argparse
+import datetime
+import random
+import os
+import logging
+from omegaconf import OmegaConf
+
+import torch
+
+import diffusers
+from diffusers import AutoencoderKL, DDIMScheduler
+
+from transformers import CLIPTextModel, CLIPTokenizer
+
+from consisti2v.models.videoldm_unet import VideoLDMUNet3DConditionModel
+from consisti2v.pipelines.pipeline_autoregress_animation import AutoregressiveAnimationPipeline
+from consisti2v.utils.util import save_videos_grid
+from diffusers.utils.import_utils import is_xformers_available
+
+def main(args, config):
+    logging.basicConfig(
+        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+        datefmt="%m/%d/%Y %H:%M:%S",
+        level=logging.INFO,
+    )
+    diffusers.utils.logging.set_verbosity_info()
+
+    time_str = datetime.datetime.now().strftime("%Y-%m-%dT%H-%M-%S")
+    savedir = f"{config.output_dir}/{config.output_name}-{time_str}"
+    os.makedirs(savedir)
+
+    samples = []
+    sample_idx = 0
+
+    ### >>> create validation pipeline >>> ###
+    if config.pipeline_pretrained_path is None:
+        noise_scheduler = DDIMScheduler(**OmegaConf.to_container(config.noise_scheduler_kwargs))
+        tokenizer       = CLIPTokenizer.from_pretrained(config.pretrained_model_path, subfolder="tokenizer", use_safetensors=True)
+        text_encoder    = CLIPTextModel.from_pretrained(config.pretrained_model_path, subfolder="text_encoder")
+        vae             = AutoencoderKL.from_pretrained(config.pretrained_model_path, subfolder="vae", use_safetensors=True)            
+        unet            = VideoLDMUNet3DConditionModel.from_pretrained(
+            config.pretrained_model_path,
+            subfolder="unet",
+            variant=config.unet_additional_kwargs['variant'],
+            temp_pos_embedding=config.unet_additional_kwargs['temp_pos_embedding'],
+            augment_temporal_attention=config.unet_additional_kwargs['augment_temporal_attention'],
+            use_temporal=True,
+            n_frames=config.sampling_kwargs['n_frames'],
+            n_temp_heads=config.unet_additional_kwargs['n_temp_heads'],
+            first_frame_condition_mode=config.unet_additional_kwargs['first_frame_condition_mode'],
+            use_frame_stride_condition=config.unet_additional_kwargs['use_frame_stride_condition'],
+            use_safetensors=True
+        )
+
+        params_unet = [p.numel() for n, p in unet.named_parameters()]
+        params_vae = [p.numel() for n, p in vae.named_parameters()]
+        params_text_encoder = [p.numel() for n, p in text_encoder.named_parameters()]
+        params = params_unet + params_vae + params_text_encoder
+        print(f"### UNet Parameters: {sum(params) / 1e6} M")
+
+        # 1. unet ckpt
+        if config.unet_path is not None:
+            if os.path.isdir(config.unet_path):
+                unet_dict = VideoLDMUNet3DConditionModel.from_pretrained(config.unet_path)
+                m, u = unet.load_state_dict(unet_dict.state_dict(), strict=False)
+                assert len(u) == 0
+                del unet_dict
+            else:
+                checkpoint_dict = torch.load(config.unet_path, map_location="cpu")
+                state_dict = checkpoint_dict["state_dict"] if "state_dict" in checkpoint_dict else checkpoint_dict
+                if config.unet_ckpt_prefix is not None:
+                    state_dict = {k.replace(config.unet_ckpt_prefix, ''): v for k, v in state_dict.items()}
+                m, u = unet.load_state_dict(state_dict, strict=False)
+                assert len(u) == 0
+
+        if is_xformers_available() and int(torch.__version__.split(".")[0]) < 2:
+            unet.enable_xformers_memory_efficient_attention()
+
+        pipeline = AutoregressiveAnimationPipeline(
+            vae=vae, text_encoder=text_encoder, tokenizer=tokenizer, unet=unet, scheduler=noise_scheduler)
+    
+    else:
+        pipeline = AutoregressiveAnimationPipeline.from_pretrained(config.pipeline_pretrained_path)
+
+    pipeline.to("cuda")
+
+    # (frameinit) initialize frequency filter for noise reinitialization -------------
+    if config.frameinit_kwargs.enable:
+        pipeline.init_filter(
+            width         = config.sampling_kwargs.width,
+            height        = config.sampling_kwargs.height,
+            video_length  = config.sampling_kwargs.n_frames,
+            filter_params = config.frameinit_kwargs.filter_params,
+        )
+    # -------------------------------------------------------------------------------
+    ### <<< create validation pipeline <<< ###
+
+    if args.prompt is not None:
+        prompts = [args.prompt]
+        n_prompts = [args.n_prompt]
+        first_frame_paths = [args.path_to_first_frame]
+        random_seeds = [int(args.seed)] if args.seed != "random" else "random"
+    else:
+        prompt_config = OmegaConf.load(args.prompt_config)
+        prompts = prompt_config.prompts
+        n_prompts = list(prompt_config.n_prompts) * len(prompts) if len(prompt_config.n_prompts) == 1 else prompt_config.n_prompts
+        first_frame_paths = prompt_config.path_to_first_frames
+        random_seeds = prompt_config.seeds
+    
+    if random_seeds == "random":
+        random_seeds = [random.randint(0, 1e5) for _ in range(len(prompts))]
+    else:
+        random_seeds = [random_seeds] if isinstance(random_seeds, int) else list(random_seeds)
+        random_seeds = random_seeds * len(prompts) if len(random_seeds) == 1 else random_seeds
+    
+    config.prompt_kwargs = OmegaConf.create({"random_seeds": [], "prompts": prompts, "n_prompts": n_prompts, "first_frame_paths": first_frame_paths})
+    for prompt_idx, (prompt, n_prompt, first_frame_path, random_seed) in enumerate(zip(prompts, n_prompts, first_frame_paths, random_seeds)):
+        # manually set random seed for reproduction
+        if random_seed != -1: torch.manual_seed(random_seed)
+        else: torch.seed()
+        config.prompt_kwargs.random_seeds.append(torch.initial_seed())
+        
+        print(f"current seed: {torch.initial_seed()}")
+        print(f"sampling {prompt} ...")
+        sample = pipeline(
+            prompt,
+            negative_prompt       = n_prompt,
+            first_frame_paths     = first_frame_path,
+            num_inference_steps   = config.sampling_kwargs.steps,
+            guidance_scale_txt    = config.sampling_kwargs.guidance_scale_txt,
+            guidance_scale_img    = config.sampling_kwargs.guidance_scale_img,
+            width                 = config.sampling_kwargs.width,
+            height                = config.sampling_kwargs.height,
+            video_length          = config.sampling_kwargs.n_frames,
+            noise_sampling_method = config.unet_additional_kwargs['noise_sampling_method'],
+            noise_alpha           = float(config.unet_additional_kwargs['noise_alpha']),
+            eta                   = config.sampling_kwargs.ddim_eta,
+            frame_stride          = config.sampling_kwargs.frame_stride,
+            guidance_rescale      = config.sampling_kwargs.guidance_rescale,
+            num_videos_per_prompt = config.sampling_kwargs.num_videos_per_prompt,
+            autoregress_steps     = config.sampling_kwargs.autoregress_steps,
+            use_frameinit          = config.frameinit_kwargs.enable,
+            frameinit_noise_level  = config.frameinit_kwargs.noise_level,
+        ).videos
+        samples.append(sample)
+
+        prompt = "-".join((prompt.replace("/", "").split(" ")[:10])).replace(":", "")
+        if sample.shape[0] > 1:
+            for cnt, samp in enumerate(sample):
+                save_videos_grid(samp.unsqueeze(0), f"{savedir}/sample/{sample_idx}-{cnt + 1}-{prompt}.{args.format}", format=args.format)
+        else:
+            save_videos_grid(sample, f"{savedir}/sample/{sample_idx}-{prompt}.{args.format}", format=args.format)
+        print(f"save to {savedir}/sample/{prompt}.{args.format}")
+        
+        sample_idx += 1
+
+    samples = torch.concat(samples)
+    save_videos_grid(samples, f"{savedir}/sample.{args.format}", n_rows=4, format=args.format)
+
+    OmegaConf.save(config, f"{savedir}/config.yaml")
+
+    if args.save_model:
+        pipeline.save_pretrained(f"{savedir}/model")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--inference_config", type=str, default="configs/inference/inference_autoregress.yaml")
+    parser.add_argument("--prompt", "-p", type=str, default=None)
+    parser.add_argument("--n_prompt", "-n", type=str, default="")
+    parser.add_argument("--seed", type=str, default="random")
+    parser.add_argument("--path_to_first_frame", "-f", type=str, default=None)
+    parser.add_argument("--prompt_config", type=str, default="configs/prompts/default.yaml")
+    parser.add_argument("--format", type=str, default="gif", choices=["gif", "mp4"])
+    parser.add_argument("--save_model", action="store_true")
+    parser.add_argument("optional_args", nargs='*', default=[])
+    args = parser.parse_args()
+
+    config = OmegaConf.load(args.inference_config)
+
+    if args.optional_args:
+        modified_config = OmegaConf.from_dotlist(args.optional_args)
+        config = OmegaConf.merge(config, modified_config)
+
+    main(args, config)

train.py

@@ -0,0 +1,617 @@
+import os
+import math
+import wandb
+import random
+import time
+import logging
+import inspect
+import argparse
+import datetime
+import numpy as np
+
+from pathlib import Path
+from tqdm.auto import tqdm
+from einops import rearrange, repeat
+from omegaconf import OmegaConf
+from typing import Dict, Optional, Tuple
+
+import torch
+import torch.nn.functional as F
+
+import diffusers
+from diffusers import AutoencoderKL, DDIMScheduler
+from diffusers.optimization import get_scheduler
+from diffusers.utils import check_min_version
+from diffusers.utils.import_utils import is_xformers_available
+from diffusers.training_utils import EMAModel
+
+import transformers
+from transformers import CLIPTextModel, CLIPTokenizer
+
+from accelerate import Accelerator, DistributedDataParallelKwargs, InitProcessGroupKwargs
+from accelerate.logging import get_logger
+from accelerate.utils import set_seed
+
+from consisti2v.data.dataset import WebVid10M, Pexels, JointDataset
+from consisti2v.models.videoldm_unet import VideoLDMUNet3DConditionModel
+from consisti2v.pipelines.pipeline_conditional_animation import ConditionalAnimationPipeline
+from consisti2v.utils.util import save_videos_grid
+
+logger = get_logger(__name__, log_level="INFO")
+
+def main(
+    name: str,
+    use_wandb: bool,
+
+    is_image: bool,
+    
+    output_dir: str,
+    pretrained_model_path: str,
+
+    train_data: Dict,
+    validation_data: Dict,
+
+    cfg_random_null_text_ratio: float = 0.1,
+    cfg_random_null_img_ratio: float = 0.0,
+    
+    resume_from_checkpoint: Optional[str] = None,
+    unet_additional_kwargs: Dict = {},
+    use_ema: bool = False,
+    ema_decay: float = 0.9999,
+    noise_scheduler_kwargs = None,
+    
+    max_train_epoch: int = -1,
+    max_train_steps: int = 100,
+    validation_steps: int = 100,
+
+    learning_rate: float = 3e-5,
+    scale_lr: bool = False,
+    lr_warmup_steps: int = 0,
+    lr_scheduler: str = "constant",
+
+    trainable_modules: Tuple[str] = (None, ),
+    num_workers: int = 32,
+    train_batch_size: int = 1,
+    adam_beta1: float = 0.9,
+    adam_beta2: float = 0.999,
+    adam_weight_decay: float = 1e-2,
+    adam_epsilon: float = 1e-08,
+    max_grad_norm: float = 1.0,
+    gradient_accumulation_steps: int = 1,
+    gradient_checkpointing: bool = False,
+    checkpointing_epochs: int = 5,
+    checkpointing_steps: int = -1,
+
+    mixed_precision: Optional[str] = "fp16",
+    enable_xformers_memory_efficient_attention: bool = True,
+
+    seed: Optional[int] = 42,
+    is_debug: bool = False,
+):
+    check_min_version("0.10.0.dev0")
+    *_, config = inspect.getargvalues(inspect.currentframe())
+    config = {k: v for k, v in config.items() if k != 'config' and k != '_'}
+
+    ddp_kwargs = DistributedDataParallelKwargs(find_unused_parameters=True if not is_image else False)
+    init_kwargs = InitProcessGroupKwargs(timeout=datetime.timedelta(seconds=3600))
+
+    accelerator = Accelerator(
+        gradient_accumulation_steps=gradient_accumulation_steps,
+        mixed_precision=mixed_precision,
+        kwargs_handlers=[ddp_kwargs, init_kwargs],
+    )
+    
+    if seed is not None:
+        set_seed(seed)
+
+    # Logging folder
+    folder_name = "debug" if is_debug else name + datetime.datetime.now().strftime("-%Y-%m-%dT%H-%M-%S")
+    output_dir = os.path.join(output_dir, folder_name)
+    if is_debug and os.path.exists(output_dir):
+        os.system(f"rm -rf {output_dir}")
+
+    # Make one log on every process with the configuration for debugging.
+    logging.basicConfig(
+        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+        datefmt="%m/%d/%Y %H:%M:%S",
+        level=logging.INFO,
+    )
+    logger.info(accelerator.state, main_process_only=False)
+
+    if accelerator.is_local_main_process:
+        transformers.utils.logging.set_verbosity_warning()
+        diffusers.utils.logging.set_verbosity_info()
+    else:
+        transformers.utils.logging.set_verbosity_error()
+        diffusers.utils.logging.set_verbosity_error()
+
+    if accelerator.is_main_process and (not is_debug) and use_wandb:
+        project_name = "text_image_to_video" if not is_image else "image_finetune"
+        wandb.init(project=project_name, name=folder_name, config=config)
+    accelerator.wait_for_everyone()
+
+    # Handle the output folder creation
+    if accelerator.is_main_process:
+        os.makedirs(output_dir, exist_ok=True)
+        os.makedirs(f"{output_dir}/samples", exist_ok=True)
+        os.makedirs(f"{output_dir}/sanity_check", exist_ok=True)
+        os.makedirs(f"{output_dir}/checkpoints", exist_ok=True)
+        OmegaConf.save(config, os.path.join(output_dir, 'config.yaml'))
+
+    # TODO: change all datasets to fps+duration in the future
+    if train_data.dataset == "pexels":
+        train_data.sample_n_frames = train_data.sample_duration * train_data.sample_fps
+    elif train_data.dataset == "joint":
+        if train_data.sample_duration is not None:
+            train_data.sample_n_frames = train_data.sample_duration * train_data.sample_fps
+    # Load scheduler, tokenizer and models.
+    noise_scheduler = DDIMScheduler(**OmegaConf.to_container(noise_scheduler_kwargs))
+    vae             = AutoencoderKL.from_pretrained(pretrained_model_path, subfolder="vae")
+    tokenizer       = CLIPTokenizer.from_pretrained(pretrained_model_path, subfolder="tokenizer")
+    text_encoder    = CLIPTextModel.from_pretrained(pretrained_model_path, subfolder="text_encoder")
+    unet            = VideoLDMUNet3DConditionModel.from_pretrained(
+        pretrained_model_path,
+        subfolder="unet",
+        variant=unet_additional_kwargs['variant'],
+        use_temporal=True if not is_image else False,
+        temp_pos_embedding=unet_additional_kwargs['temp_pos_embedding'],
+        augment_temporal_attention=unet_additional_kwargs['augment_temporal_attention'],
+        n_frames=train_data.sample_n_frames if not is_image else 2,
+        n_temp_heads=unet_additional_kwargs['n_temp_heads'],
+        first_frame_condition_mode=unet_additional_kwargs['first_frame_condition_mode'],
+        use_frame_stride_condition=unet_additional_kwargs['use_frame_stride_condition'],
+        use_safetensors=True
+    )
+
+    # Freeze vae and text_encoder
+    vae.requires_grad_(False)
+    text_encoder.requires_grad_(False)
+    unet.train()
+
+    if use_ema:
+        ema_unet = VideoLDMUNet3DConditionModel.from_pretrained(
+            pretrained_model_path,
+            subfolder="unet",
+            variant=unet_additional_kwargs['variant'],
+            use_temporal=True if not is_image else False,
+            temp_pos_embedding=unet_additional_kwargs['temp_pos_embedding'],
+            augment_temporal_attention=unet_additional_kwargs['augment_temporal_attention'],
+            n_frames=train_data.sample_n_frames if not is_image else 2,
+            n_temp_heads=unet_additional_kwargs['n_temp_heads'],
+            first_frame_condition_mode=unet_additional_kwargs['first_frame_condition_mode'],
+            use_frame_stride_condition=unet_additional_kwargs['use_frame_stride_condition'],
+            use_safetensors=True
+        )
+        ema_unet = EMAModel(ema_unet.parameters(), decay=ema_decay, model_cls=VideoLDMUNet3DConditionModel, model_config=ema_unet.config)
+    
+    # Set unet trainable parameters
+    train_all_parameters = False
+    for trainable_module_name in trainable_modules:
+        if trainable_module_name == 'all':
+            unet.requires_grad_(True)
+            train_all_parameters = True
+            break
+
+    if not train_all_parameters:
+        unet.requires_grad_(False)
+        for name, param in unet.named_parameters():
+            for trainable_module_name in trainable_modules:
+                if trainable_module_name in name:
+                    param.requires_grad = True
+                    break
+
+    # Enable xformers
+    if enable_xformers_memory_efficient_attention and int(torch.__version__.split(".")[0]) < 2:
+        if is_xformers_available():
+            unet.enable_xformers_memory_efficient_attention()
+        else:
+            raise ValueError("xformers is not available. Make sure it is installed correctly")
+
+    def save_model_hook(models, weights, output_dir):
+        if accelerator.is_main_process:
+            if use_ema:
+                ema_unet.save_pretrained(os.path.join(output_dir, "unet_ema"))
+
+            for i, model in enumerate(models):
+                model.save_pretrained(os.path.join(output_dir, "unet"))
+
+                # make sure to pop weight so that corresponding model is not saved again
+                weights.pop()
+
+    def load_model_hook(models, input_dir):
+        if use_ema:
+            load_model = EMAModel.from_pretrained(os.path.join(input_dir, "unet_ema"), VideoLDMUNet3DConditionModel)
+            ema_unet.load_state_dict(load_model.state_dict())
+            ema_unet.to(accelerator.device)
+            del load_model
+
+        for i in range(len(models)):
+            # pop models so that they are not loaded again
+            model = models.pop()
+
+            # load diffusers style into model
+            load_model = VideoLDMUNet3DConditionModel.from_pretrained(input_dir, subfolder="unet")
+            model.register_to_config(**load_model.config)
+
+            model.load_state_dict(load_model.state_dict())
+            del load_model
+
+    accelerator.register_save_state_pre_hook(save_model_hook)
+    accelerator.register_load_state_pre_hook(load_model_hook)
+
+    # Enable gradient checkpointing
+    if gradient_checkpointing:
+        unet.enable_gradient_checkpointing()
+
+    if scale_lr:
+        learning_rate = (learning_rate * gradient_accumulation_steps * train_batch_size * accelerator.num_processes)
+    
+    trainable_params = list(filter(lambda p: p.requires_grad, unet.parameters()))
+    optimizer = torch.optim.AdamW(
+        trainable_params,
+        lr=learning_rate,
+        betas=(adam_beta1, adam_beta2),
+        weight_decay=adam_weight_decay,
+        eps=adam_epsilon,
+    )
+
+    logger.info(f"trainable params number: {len(trainable_params)}")
+    logger.info(f"trainable params scale: {sum(p.numel() for p in trainable_params) / 1e6:.3f} M")
+
+    # Get the training dataset
+    if train_data['dataset'] == "webvid":
+        train_dataset = WebVid10M(**train_data, is_image=is_image)
+    elif train_data['dataset'] == "pexels":
+        train_dataset = Pexels(**train_data, is_image=is_image)
+    elif train_data['dataset'] == "joint":
+        train_dataset = JointDataset(**train_data, is_image=is_image)
+    else:
+        raise ValueError(f"Unknown dataset {train_data['dataset']}")
+
+    # DataLoaders creation:
+    train_dataloader = torch.utils.data.DataLoader(
+        train_dataset,
+        shuffle=True,
+        batch_size=train_batch_size,
+        num_workers=num_workers,
+        pin_memory=True,
+    )
+
+    # Get the training iteration
+    if max_train_steps == -1:
+        assert max_train_epoch != -1
+        max_train_steps = max_train_epoch * len(train_dataloader)
+        
+    if checkpointing_steps == -1:
+        assert checkpointing_epochs != -1
+        checkpointing_steps = checkpointing_epochs * len(train_dataloader)
+
+    # Scheduler
+    lr_scheduler = get_scheduler(
+        lr_scheduler,
+        optimizer=optimizer,
+        num_warmup_steps=lr_warmup_steps * gradient_accumulation_steps,
+        num_training_steps=max_train_steps * gradient_accumulation_steps,
+    )
+
+    # Validation pipeline
+    validation_pipeline = ConditionalAnimationPipeline(
+        unet=unet, vae=vae, tokenizer=tokenizer, text_encoder=text_encoder, scheduler=noise_scheduler,
+    )
+    validation_pipeline.enable_vae_slicing()
+
+    # Prepare everything with our `accelerator`.
+    unet, optimizer, train_dataloader, lr_scheduler = accelerator.prepare(
+        unet, optimizer, train_dataloader, lr_scheduler
+    )
+
+    # For mixed precision training we cast the text_encoder and vae weights to half-precision
+    # as these models are only used for inference, keeping weights in full precision is not required.
+    weight_dtype = torch.float32
+    if accelerator.mixed_precision == "fp16":
+        weight_dtype = torch.float16
+    elif accelerator.mixed_precision == "bf16":
+        weight_dtype = torch.bfloat16
+
+    if use_ema:
+        ema_unet.to(accelerator.device)
+
+    # Move text_encode and vae to gpu and cast to weight_dtype
+    text_encoder.to(accelerator.device, dtype=weight_dtype)
+    vae.to(accelerator.device, dtype=weight_dtype)
+
+    # We need to recalculate our total training steps as the size of the training dataloader may have changed.
+    num_update_steps_per_epoch = math.ceil(len(train_dataloader) / gradient_accumulation_steps)
+    # Afterwards we recalculate our number of training epochs
+    num_train_epochs = math.ceil(max_train_steps / num_update_steps_per_epoch)
+
+    # Train!
+    total_batch_size = train_batch_size * accelerator.num_processes * gradient_accumulation_steps
+
+    logger.info("***** Running training *****")
+    logger.info(f"  Num examples = {len(train_dataset)}")
+    logger.info(f"  Num Epochs = {num_train_epochs}")
+    logger.info(f"  Instantaneous batch size per device = {train_batch_size}")
+    logger.info(f"  Total train batch size (w. parallel, distributed & accumulation) = {total_batch_size}")
+    logger.info(f"  Gradient Accumulation steps = {gradient_accumulation_steps}")
+    logger.info(f"  Total optimization steps = {max_train_steps}")
+
+    global_step = 0
+    first_epoch = 0
+
+    # Load pretrained unet weights
+    if resume_from_checkpoint is not None:
+        logger.info(f"Resuming from checkpoint: {resume_from_checkpoint}")
+        accelerator.load_state(resume_from_checkpoint)
+        global_step = int(resume_from_checkpoint.split("-")[-1])
+
+        initial_global_step = global_step
+        first_epoch = global_step // num_update_steps_per_epoch
+        logger.info(f"global_step: {global_step}")
+        logger.info(f"first_epoch: {first_epoch}")
+    else:
+        initial_global_step = 0
+
+    # Only show the progress bar once on each machine.
+    progress_bar = tqdm(range(0, max_train_steps), initial=initial_global_step, desc="Steps", disable=not accelerator.is_main_process)
+
+    for epoch in range(first_epoch, num_train_epochs):
+        train_loss = 0.0
+        train_grad_norm = 0.0
+        data_loading_time = 0.0
+        prepare_everything_time = 0.0
+        network_forward_time = 0.0
+        network_backward_time = 0.0
+
+        t0 = time.time()
+        for step, batch in enumerate(train_dataloader):
+            t1 = time.time()
+            if cfg_random_null_text_ratio > 0.0:
+                batch['text'] = [name if random.random() > cfg_random_null_text_ratio else "" for name in batch['text']]
+                
+            # Data batch sanity check
+            if accelerator.is_main_process and epoch == first_epoch and step == 0:
+                pixel_values, texts = batch['pixel_values'].cpu(), batch['text']
+                pixel_values = rearrange(pixel_values, "b f c h w -> b c f h w")
+                for idx, (pixel_value, text) in enumerate(zip(pixel_values, texts)):
+                    pixel_value = pixel_value[None, ...]
+                    save_videos_grid(pixel_value, f"{output_dir}/sanity_check/{'-'.join(text.replace('/', '').split()[:10]) if not text == '' else f'no_text-{idx}'}.gif", rescale=True)
+                    
+            ### >>>> Training >>>> ###
+            with accelerator.accumulate(unet):
+                # Convert videos to latent space            
+                pixel_values = batch["pixel_values"].to(weight_dtype)
+                video_length = pixel_values.shape[1]
+                pixel_values = rearrange(pixel_values, "b f c h w -> (b f) c h w")
+                latents = vae.encode(pixel_values).latent_dist
+                latents = latents.sample()
+                latents = rearrange(latents, "(b f) c h w -> b c f h w", f=video_length)
+
+                latents = latents * vae.config.scaling_factor
+
+                if unet_additional_kwargs["first_frame_condition_mode"] != "none":
+                    # Get first frame latents
+                    first_frame_latents = latents[:, :, 0:1, :, :]
+
+                # Sample noise that we'll add to the latents
+                if unet_additional_kwargs['noise_sampling_method'] == 'vanilla':
+                    noise = torch.randn_like(latents)
+                elif unet_additional_kwargs['noise_sampling_method'] == 'pyoco_mixed':
+                    noise_alpha_squared = float(unet_additional_kwargs['noise_alpha']) ** 2
+                    shared_noise = torch.randn_like(latents[:, :, 0:1, :, :]) * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared))
+                    ind_noise = torch.randn_like(latents) * math.sqrt(1 / (1 + noise_alpha_squared))
+                    noise = shared_noise + ind_noise
+                elif unet_additional_kwargs['noise_sampling_method'] == 'pyoco_progressive':
+                    noise_alpha_squared = float(unet_additional_kwargs['noise_alpha']) ** 2
+                    noise = torch.randn_like(latents)
+                    ind_noise = torch.randn_like(latents) * math.sqrt(1 / (1 + noise_alpha_squared))
+                    for i in range(1, noise.shape[2]):
+                        noise[:, :, i, :, :] = noise[:, :, i - 1, :, :] * math.sqrt((noise_alpha_squared) / (1 + noise_alpha_squared)) + ind_noise[:, :, i, :, :]
+                else:
+                    raise ValueError(f"Unknown noise sampling method {unet_additional_kwargs['noise_sampling_method']}")
+
+                bsz = latents.shape[0]
+            
+                # Sample a random timestep for each video
+                timesteps = torch.randint(0, noise_scheduler.config.num_train_timesteps, (bsz,), device=latents.device)
+                timesteps = timesteps.long()
+            
+                # Add noise to the latents according to the noise magnitude at each timestep
+                # (this is the forward diffusion process)
+                noisy_latents = noise_scheduler.add_noise(latents, noise, timesteps)
+
+                if cfg_random_null_img_ratio > 0.0:
+                    for i in range(first_frame_latents.shape[0]):
+                        if random.random() <= cfg_random_null_img_ratio:
+                            first_frame_latents[i, :, :, :, :] = noisy_latents[i, :, 0:1, :, :]
+
+                # Remove the first noisy latent from the latents if we're conditioning on the first frame
+                if unet_additional_kwargs["first_frame_condition_mode"] != "none":
+                    noisy_latents = noisy_latents[:, :, 1:, :, :]
+            
+                # Get the text embedding for conditioning
+                prompt_ids = tokenizer(
+                    batch['text'], max_length=tokenizer.model_max_length, padding="max_length", truncation=True, return_tensors="pt"
+                ).input_ids.to(latents.device)
+                encoder_hidden_states = text_encoder(prompt_ids)[0]
+                
+                # Get the target for loss depending on the prediction type
+                if noise_scheduler.config.prediction_type == "epsilon":
+                    target = noise
+                elif noise_scheduler.config.prediction_type == "v_prediction":
+                    target = noise_scheduler.get_velocity(latents, noise, timesteps)
+                else:
+                    raise ValueError(f"Unknown prediction type {noise_scheduler.config.prediction_type}")
+
+                timesteps = repeat(timesteps, "b -> b f", f=video_length)
+                timesteps = rearrange(timesteps, "b f -> (b f)")
+
+                frame_stride = None
+                if unet_additional_kwargs["use_frame_stride_condition"]:
+                    frame_stride = batch['stride'].to(latents.device)
+                    frame_stride = frame_stride.long()
+                    frame_stride = repeat(frame_stride, "b -> b f", f=video_length)
+                    frame_stride = rearrange(frame_stride, "b f -> (b f)")
+
+                t2 = time.time()
+
+                # Predict the noise residual and compute loss
+                if unet_additional_kwargs["first_frame_condition_mode"] != "none":
+                    model_pred = unet(noisy_latents, timesteps, encoder_hidden_states, first_frame_latents=first_frame_latents, frame_stride=frame_stride).sample
+                    loss = F.mse_loss(model_pred.float(), target.float()[:, :, 1:, :, :], reduction="mean")
+                else:
+                    model_pred = unet(noisy_latents, timesteps, encoder_hidden_states).sample
+                    loss = F.mse_loss(model_pred.float(), target.float(), reduction="mean")
+                
+                t3 = time.time()
+                
+                avg_loss = accelerator.gather(loss.repeat(train_batch_size)).mean()
+                train_loss += avg_loss.item() / gradient_accumulation_steps
+
+                # Backpropagate
+                accelerator.backward(loss)
+                if accelerator.sync_gradients:
+                    grad_norm = accelerator.clip_grad_norm_(unet.parameters(), max_grad_norm)
+                    avg_grad_norm = accelerator.gather(grad_norm.repeat(train_batch_size)).mean()
+                    train_grad_norm += avg_grad_norm.item() / gradient_accumulation_steps
+
+                optimizer.step()
+                lr_scheduler.step()
+                optimizer.zero_grad()
+
+                t4 = time.time()
+
+                data_loading_time += (t1 - t0) / gradient_accumulation_steps
+                prepare_everything_time += (t2 - t1) / gradient_accumulation_steps
+                network_forward_time += (t3 - t2) / gradient_accumulation_steps
+                network_backward_time += (t4 - t3) / gradient_accumulation_steps
+
+                t0 = time.time()
+            
+            ### <<<< Training <<<< ###
+            
+            # Checks if the accelerator has performed an optimization step behind the scenes
+            if accelerator.sync_gradients:
+                if use_ema:
+                    ema_unet.step(unet.parameters())
+                progress_bar.update(1)
+                global_step += 1
+
+                # Wandb logging
+                if accelerator.is_main_process and (not is_debug) and use_wandb:
+                    wandb.log({"metrics/train_loss": train_loss}, step=global_step)
+                    wandb.log({"metrics/train_grad_norm": train_grad_norm}, step=global_step)
+                    
+                    wandb.log({"profiling/train_data_loading_time": data_loading_time}, step=global_step)
+                    wandb.log({"profiling/train_prepare_everything_time": prepare_everything_time}, step=global_step)
+                    wandb.log({"profiling/train_network_forward_time": network_forward_time}, step=global_step)
+                    wandb.log({"profiling/train_network_backward_time": network_backward_time}, step=global_step)
+                    # accelerator.log({"train_loss": train_loss}, step=global_step)
+                train_loss = 0.0
+                train_grad_norm = 0.0
+                data_loading_time = 0.0
+                prepare_everything_time = 0.0
+                network_forward_time = 0.0
+                network_backward_time = 0.0
+                
+                # Save checkpoint
+                if global_step % checkpointing_steps == 0:
+                    if accelerator.is_main_process:
+                        save_path = os.path.join(output_dir, f"checkpoints/checkpoint-{global_step}")
+                        accelerator.save_state(save_path)
+                        logger.info(f"Saved state to {save_path} (global_step: {global_step})")
+                
+                # Periodically validation
+                if accelerator.is_main_process and global_step % validation_steps == 0:
+                    if use_ema:
+                        # Store the UNet parameters temporarily and load the EMA parameters to perform inference.
+                        ema_unet.store(unet.parameters())
+                        ema_unet.copy_to(unet.parameters())
+
+                    samples = []
+                    wandb_samples = []
+                    
+                    generator = torch.Generator(device=latents.device)
+                    generator.manual_seed(seed)
+                    
+                    height = train_data.sample_size[0] if not isinstance(train_data.sample_size, int) else train_data.sample_size
+                    width  = train_data.sample_size[1] if not isinstance(train_data.sample_size, int) else train_data.sample_size
+
+                    prompts = validation_data.prompts
+                    
+                    first_frame_paths = [None] * len(prompts)
+                    if unet_additional_kwargs["first_frame_condition_mode"] != "none":
+                        first_frame_paths = validation_data.path_to_first_frames
+
+                    for idx, (prompt, first_frame_path) in enumerate(zip(prompts, first_frame_paths)):
+                        sample = validation_pipeline(
+                            prompt,
+                            generator    = generator,
+                            video_length = train_data.sample_n_frames if not is_image else 2,
+                            height       = height,
+                            width        = width,
+                            first_frame_paths = first_frame_path,
+                            noise_sampling_method = unet_additional_kwargs['noise_sampling_method'],
+                            noise_alpha = float(unet_additional_kwargs['noise_alpha']),
+                            **validation_data,
+                        ).videos
+                        save_videos_grid(sample, f"{output_dir}/samples/sample-{global_step}/{idx}.gif")
+                        samples.append(sample)
+                        
+                        numpy_sample = (sample.squeeze(0).permute(1, 0, 2, 3) * 255).cpu().numpy().astype(np.uint8)
+                        wandb_video = wandb.Video(numpy_sample, fps=8, caption=prompt)
+                        wandb_samples.append(wandb_video)
+                    
+                    if (not is_debug) and use_wandb:
+                        val_title = 'val_videos'
+                        wandb.log({val_title: wandb_samples}, step=global_step)
+                    
+                    samples = torch.concat(samples)
+                    save_path = f"{output_dir}/samples/sample-{global_step}.gif"
+                    save_videos_grid(samples, save_path)
+
+                    logger.info(f"Saved samples to {save_path}")
+
+                    if use_ema:
+                        # Switch back to the original UNet parameters.
+                        ema_unet.restore(unet.parameters())
+                
+            logs = {"step_loss": loss.detach().item(), "lr": lr_scheduler.get_last_lr()[0]}
+            progress_bar.set_postfix(**logs)
+            
+            if accelerator.is_main_process and (not is_debug) and use_wandb:
+                wandb.log({"metrics/train_lr": lr_scheduler.get_last_lr()[0]}, step=global_step)
+            
+            if global_step >= max_train_steps:
+                break
+            
+    # Create the pipeline using the trained modules and save it.
+    accelerator.wait_for_everyone()
+    if accelerator.is_main_process:
+        unet = accelerator.unwrap_model(unet)
+        pipeline = ConditionalAnimationPipeline(
+            text_encoder=text_encoder,
+            vae=vae,
+            unet=unet,
+            tokenizer=tokenizer,
+            scheduler=noise_scheduler,
+        )
+        pipeline.save_pretrained(f"{output_dir}/final_checkpoint")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--config",   type=str, required=True)
+    parser.add_argument("--name", "-n", type=str, default="")
+    parser.add_argument("--wandb",    action="store_true")
+    parser.add_argument("optional_args", nargs='*', default=[])
+    args = parser.parse_args()
+
+    name   = args.name + "_" + Path(args.config).stem
+    config = OmegaConf.load(args.config)
+
+    if args.optional_args:
+        modified_config = OmegaConf.from_dotlist(args.optional_args)
+        config = OmegaConf.merge(config, modified_config)
+
+    main(name=name, use_wandb=args.wandb, **config)