v2-diffusers

feature_extractor/preprocessor_config.json

@@ -1,8 +1,12 @@
 {
-  "crop_size": 224,
+  "crop_size": {
+    "height": 224,
+    "width": 224
+  },
   "do_center_crop": true,
   "do_convert_rgb": true,
   "do_normalize": true,
+  "do_rescale": true,
   "do_resize": true,
   "feature_extractor_type": "CLIPFeatureExtractor",
   "image_mean": [
@@ -10,11 +14,15 @@
     0.4578275,
     0.40821073
   ],
+  "image_processor_type": "CLIPImageProcessor",
   "image_std": [
     0.26862954,
     0.26130258,
     0.27577711
   ],
   "resample": 3,
-  "size": 224
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 224
+  }
 }

model_index.json

@@ -1,10 +1,11 @@
 {
   "_class_name": "StableDiffusionPipeline",
-  "_diffusers_version": "0.8.0.dev0",
+  "_diffusers_version": "0.10.2",
   "feature_extractor": [
     "transformers",
-    "CLIPFeatureExtractor"
+    "CLIPImageProcessor"
   ],
+  "requires_safety_checker": true,
   "safety_checker": [
     "stable_diffusion",
     "StableDiffusionSafetyChecker"

safety_checker/config.json

@@ -1,5 +1,5 @@
 {
-  "_commit_hash": "4bb648a606ef040e7685bde262611766a5fdd67b",
+  "_commit_hash": null,
   "_name_or_path": "CompVis/stable-diffusion-safety-checker",
   "architectures": [
     "StableDiffusionSafetyChecker"
@@ -62,6 +62,7 @@
     "pad_token_id": 1,
     "prefix": null,
     "problem_type": null,
+    "projection_dim": 512,
     "pruned_heads": {},
     "remove_invalid_values": false,
     "repetition_penalty": 1.0,
@@ -79,7 +80,7 @@
     "top_p": 1.0,
     "torch_dtype": null,
     "torchscript": false,
-    "transformers_version": "4.24.0",
+    "transformers_version": "4.25.1",
     "typical_p": 1.0,
     "use_bfloat16": false,
     "vocab_size": 49408
@@ -90,7 +91,7 @@
     "num_attention_heads": 12,
     "num_hidden_layers": 12
   },
-  "torch_dtype": "float32",
+  "torch_dtype": "float16",
   "transformers_version": null,
   "vision_config": {
     "_name_or_path": "",
@@ -148,6 +149,7 @@
     "patch_size": 14,
     "prefix": null,
     "problem_type": null,
+    "projection_dim": 512,
     "pruned_heads": {},
     "remove_invalid_values": false,
     "repetition_penalty": 1.0,
@@ -165,7 +167,7 @@
     "top_p": 1.0,
     "torch_dtype": null,
     "torchscript": false,
-    "transformers_version": "4.24.0",
+    "transformers_version": "4.25.1",
     "typical_p": 1.0,
     "use_bfloat16": false
   },

text_encoder/pytorch_model.bin → safety_checker/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:90cacd05ab6f5dc1989fea6b114897494f6f0c3a6a00399d4fce4210ff6f066b
-size 492307041
+oid sha256:69c4f8737c7620a738704911a1c7be4dce7bad37bead09e9e499da8810ffb26d
+size 608018446

scheduler/scheduler_config.json

@@ -1,11 +1,12 @@
 {
   "_class_name": "PNDMScheduler",
-  "_diffusers_version": "0.8.0.dev0",
+  "_diffusers_version": "0.10.2",
   "beta_end": 0.012,
   "beta_schedule": "scaled_linear",
   "beta_start": 0.00085,
   "clip_sample": false,
   "num_train_timesteps": 1000,
+  "prediction_type": "epsilon",
   "set_alpha_to_one": false,
   "skip_prk_steps": true,
   "steps_offset": 1,

text_encoder/config.json

@@ -19,7 +19,7 @@
   "num_hidden_layers": 12,
   "pad_token_id": 1,
   "projection_dim": 768,
-  "torch_dtype": "float32",
-  "transformers_version": "4.24.0",
+  "torch_dtype": "float16",
+  "transformers_version": "4.25.1",
   "vocab_size": 49408
 }

vae/diffusion_pytorch_model.bin → text_encoder/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:36bb8e1b54aba3a0914eb35fba13dcb107e9f18d379d1df2158732cd4bf56a94
-size 334711857
+oid sha256:194dbd7d114894558b4223ebbec17b285693c6bc961daad53ee4cff7be1449d5
+size 246144867

unet/config.json

@@ -1,6 +1,7 @@
 {
   "_class_name": "UNet2DConditionModel",
-  "_diffusers_version": "0.8.0.dev0",
+  "_diffusers_version": "0.10.2",
+  "_name_or_path": "./unet",
   "act_fn": "silu",
   "attention_head_dim": 8,
   "block_out_channels": [
@@ -18,6 +19,7 @@
     "DownBlock2D"
   ],
   "downsample_padding": 1,
+  "dual_cross_attention": false,
   "flip_sin_to_cos": true,
   "freq_shift": 0,
   "in_channels": 4,
@@ -25,12 +27,16 @@
   "mid_block_scale_factor": 1,
   "norm_eps": 1e-05,
   "norm_num_groups": 32,
+  "num_class_embeds": null,
+  "only_cross_attention": false,
   "out_channels": 4,
-  "sample_size": 32,
+  "sample_size": 64,
   "up_block_types": [
     "UpBlock2D",
     "CrossAttnUpBlock2D",
     "CrossAttnUpBlock2D",
     "CrossAttnUpBlock2D"
-  ]
+  ],
+  "upcast_attention": false,
+  "use_linear_projection": false
 }

safety_checker/pytorch_model.bin → unet/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:16d28f2b37109f222cdc33620fdd262102ac32112be0352a7f77e9614b35a394
-size 1216064769
+oid sha256:cb9fdfd114ce091c10b4de5a71dbec96b32e07beae937b71429504cd4f13b969
+size 1719125271

vae/config.json

@@ -1,6 +1,7 @@
 {
   "_class_name": "AutoencoderKL",
-  "_diffusers_version": "0.8.0.dev0",
+  "_diffusers_version": "0.10.2",
+  "_name_or_path": "./vae",
   "act_fn": "silu",
   "block_out_channels": [
     128,
@@ -19,7 +20,7 @@
   "layers_per_block": 2,
   "norm_num_groups": 32,
   "out_channels": 3,
-  "sample_size": 256,
+  "sample_size": 512,
   "up_block_types": [
     "UpDecoderBlock2D",
     "UpDecoderBlock2D",

unet/diffusion_pytorch_model.bin → vae/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f9eed11e55d4f2d281f059d83a305b19ca90baac9b3179b3b65026cf18db22f3
-size 3438366373
+oid sha256:11e897bda3163f66848a4abe4613001d27502c985c84d33eb10e7387f9010a2c
+size 167335315