adds new Kazusa LoRA

kazusa/README.md

@@ -0,0 +1,51 @@
+# Kyoyama Kazusa (Blue Archive)
+杏山カズサ (ブルーアーカイブ) / 쿄야마 카즈사 (블루 아카이브) / 杏山和纱 (碧蓝档案)
+
+[**Download here.**](https://huggingface.co/khanon/lora-training/blob/main/kazusa/chara-kazusa-v1c.safetensors)
+
+## Table of Contents
+- [Preview](#preview)
+- [Usage](#usage)
+- [Training](#training)
+- [Revisions](#revisions)
+
+## Preview
+![Kazusa portrait](chara-haruka-v1c.png)
+![Kazusa preview 1](example-001-v1c-16dim.png)
+![Kazusa preview 2](example-002-v1c-16dim.png)
+![Kazusa preview 3](example-003-v1c-16dim.png)
+![Kazusa preview 4](example-004-v1c-16dim.png)
+![Kazusa preview 6](example-006-v1c-16dim.png)
+
+## Usage
+- Use any or all of the following tags to summon Kazusa: `kazusa, 1girl, animal ears, colored inner hair, halo, short hair`
+- For her normal outfit: `black choker, black jacket, black pantyhose, green sailor collar, hooded jacket, white skirt, miniskirt, hairclip, pink neckerchief, school uniform, sneakers`
+  - For a closed jacket, you might need to add `open jacket, white shirt` to the negative prompt.
+  - For raised hood: `hood up`
+- For her "trying to look cool" expression: `expressionless, blush, sweatdrop, v-shaped eyebrows, embarrassed`
+  - For selfies: `selfie, (reaching towards viewer:1.2)`
+- For eating: `eating, :t, holding food, food on face`
+- Accessories: `macaron, cake, fork`
+    
+## Training
+*Exact parameters are provided in the accompanying JSON files.*
+- Trained on a set of 280 images.
+  - 7 repeats
+  - 3 batch size, 4 epochs
+  - `(280 * 7) / 3 * 4` = 2614 steps
+  - Kazusa has a lot of good art so her dataset is much larger than usual.  This seems to have let me train for longer without overfitting.
+- 0.08 loss
+- Initially tagged with WD1.4 swin-v2 model. Tags pruned/edited for consistency.
+  - [Detailed tagging methodology here.](tagging%20methodology.md)
+- `constant_with_warmup` scheduler
+- 1.5e-5 text encoder LR
+- 1.5e-4 unet LR
+- 1e-5 optimizer LR
+- Used network_dimension 128 (same as usual) / network alpha 128 (default)
+  - Resized to 16 after training
+- Training resolution 832x832.
+- Trained without VAE.
+
+## Revisions
+- v1c (2023-03-02)
+  - Initial release.

kazusa/chara-kazusa-v1c.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54ab8df42e151b882ec674b2127e7040c769f7bf99fc0d58c7c172cacfeacb5b
+size 19007600

kazusa/lora_chara_kazusa_v1c_280i7r.json

@@ -0,0 +1,62 @@
+{
+  "pretrained_model_name_or_path": "G:/sd/repo/models/Stable-diffusion/nai-animefull-final-pruned.safetensors",
+  "v2": false,
+  "v_parameterization": false,
+  "logging_dir": "",
+  "train_data_dir": "G:/sd/training/datasets/kazusa/dataset",
+  "reg_data_dir": "",
+  "output_dir": "G:/sd/lora/trained/chara/kazusa",
+  "max_resolution": "832,832",
+  "learning_rate": "1e-5",
+  "lr_scheduler": "constant_with_warmup",
+  "lr_warmup": "5",
+  "train_batch_size": 3,
+  "epoch": "4",
+  "save_every_n_epochs": "",
+  "mixed_precision": "fp16",
+  "save_precision": "fp16",
+  "seed": "31337",
+  "num_cpu_threads_per_process": 32,
+  "cache_latents": true,
+  "caption_extension": ".txt",
+  "enable_bucket": true,
+  "gradient_checkpointing": false,
+  "full_fp16": false,
+  "no_token_padding": false,
+  "stop_text_encoder_training": 0,
+  "use_8bit_adam": false,
+  "xformers": true,
+  "save_model_as": "safetensors",
+  "shuffle_caption": true,
+  "save_state": false,
+  "resume": "",
+  "prior_loss_weight": 1.0,
+  "text_encoder_lr": "1.5e-5",
+  "unet_lr": "1.5e-4",
+  "network_dim": 128,
+  "lora_network_weights": "",
+  "color_aug": false,
+  "flip_aug": false,
+  "clip_skip": 2,
+  "gradient_accumulation_steps": 1.0,
+  "mem_eff_attn": false,
+  "output_name": "chara-kazusa-v1c-128",
+  "model_list": "",
+  "max_token_length": "150",
+  "max_train_epochs": "",
+  "max_data_loader_n_workers": "",
+  "network_alpha": 128,
+  "training_comment": "Character: `kazusa, 1girl, animal ears, short hair, colored inner hair, halo`\nStandard outfit: `school uniform, black hooded jacket, hairclip, green sailor collar, white skirt, miniskirt, black pantyhose, sneakers`\nTrying-to-look-cool expression: `blush, expressionless, sweatdrop, v-shaped eyebrows, embarrassed`\nAccessories: `holding food`, `holding fork`, `cake`, `macaron`, `fork`\n\nNot all tags are necessary.\n\n(280 images * 7 repeats) / 3 batch size * 4 epochs = 2614 steps",
+  "keep_tokens": 2,
+  "lr_scheduler_num_cycles": "",
+  "lr_scheduler_power": "",
+  "persistent_data_loader_workers": true,
+  "bucket_no_upscale": true,
+  "random_crop": false,
+  "bucket_reso_steps": 64.0,
+  "caption_dropout_every_n_epochs": 0.0,
+  "caption_dropout_rate": 0,
+  "optimizer": "AdamW8bit",
+  "optimizer_args": "",
+  "noise_offset": ""
+}

kazusa/tagging methodology.md

@@ -0,0 +1,88 @@
+Tagging methodology for Kazusa (blue archive)
+
+Start with WD1.4 Swinv2 at 0.25 confidence.
+
+- Tag unique features
+  - `halo` / `demon horns` / `low wings`
+  - Remove when not present or out of view.  WD1.4 likes putting `halo` even on images where no halo is visible.
+  - Kazusa: `halo` / `animal ears`
+    - Pruned `extra ears` as it seems redundant.
+- Tag outfit variants with a single master tag
+  - Kazusa:
+    - Uniform: `school uniform` / `black jacket`
+      - Sometimes the jacket appears without anything else, which was not tagged `school uniform`
+    - Non-canon costumes
+      - Add `alternate costume`
+  - Nudity (WD1.4 usually does this accurately)
+    - `nude` / `completely nude`
+- Prune eye colors
+  - Keep tags which describe unusual eye features (`multicolored eyes`, `heterochromia`, `slit pupils`) as they can otherwise be too subtle and inconsistently drawn for the AI to learn
+- Prune hair colors
+  - This includes `two-toned hair`, `gradiant hair`, etc.  The AI learns all of these very consistently without the tags, likely because artists tend to draw them consistently
+- Partially prune hair styles
+  - Leave key style tags like `twintails`, `ponytail`, `short hair with long locks`.
+  - Prune exceedingly common tags like `bangs` / `sidelocks` / `eyebrows visible through hair` / `hair between eyes`, etc.
+  - Prune length, except for images which differ from the character's usual length
+    - Add `alternate hairstyle` and/or `alternate hair length` for these
+  - Kazusa: `short hair, colored inner hair` -- while I would usually prune these, they're really her only defining hairstyle traits
+- Fixup hair ornaments
+  - Prune generic `hair ornament` in favor of more specificity
+    - `hairclip` / `black headband` / `hair flower` / `hair ribbon`, etc.
+  - Consolidate tags that have color variants (`headband` >> `black headband`)
+  - Kazusa: `hairclip`
+- Consolidate outfits
+  - Only tag an item when it is actually visible. If it is only barely visible along the edge of an image, keep in mind it may be cropped during bucketing.
+  - Danbooru's wiki entry for a character often provides a good list of tags commonly used to describe a character's outfits.
+  - Kazusa outfits:
+    - School Uniform
+      - `black choker`
+      - `hooded jacket`
+      - `black jacket`
+      - `green sailor collar`
+      - `pink neckerchief`
+      - `miniskirt`
+      - `pleated skirt`
+      - `white skirt`
+      - `black pantyhose`
+      - `sneakers`
+- Fixup sleeves
+  - ie. `long sleeves` / `puffy long sleeves` / `detached sleeves`
+  - You only need one, but pick one and be consistent. If sleeves aren't tagged properly the tends to add them inappropriately (such as when prompting for sleeveless outfits or nudity)
+- Fixup collars
+  - ie. `detached collar` / `collared shirt` / `choker` / etc.
+  - Same deal as sleeves, they tend to appear when unwanted if not consistently tagged according to visibility in the training data.
+- Fixup clothing state
+  - ie. `open jacket` / `open shirt` / `partially undressed` / `off shoulder`
+  - The tagger is generally pretty good at this.
+- Tag expressions
+  - WD1.4 rarely tags these, but doing them manually can help the AI reproduce a character's iconic expressions well.
+  - Start by searching for images without one of these, and add them.
+    - `open mouth`
+    - `closed mouth`
+    - `parted lips`
+  - Add less common expressions
+  - `smile` / `light smile` / `:d`
+  - `wavy mouth` / `embarrassed`
+  - `flustered` / `panicked` / `swirly eyes` / `@_@`
+  - `surprised` / `o_o` / `wide-eyed`
+  - `upset` / `annoyed` / `frustrated` / `v-shaped eyebrows`
+  - `naughty face` / `seductive smile`
+  - `smug` / `:3`
+  - `eyes closed` / `one eye closed`
+    - WD1.4 usually gets these for you.
+- Tag camera angles/composition
+  - `cowboy shot`
+  - `upper body`
+  - `full body`
+  - `portrait`
+  - `cropped torso` / `cropped legs`
+  - `feet out of frame`
+  - `from side` / `from above` / `from below` / `from behind`
+- Tag iconic poses/actions
+  - ie. `v` / `standing on one leg` 
+  - Kazusa
+    - `mouth hold`
+    - `eating`
+    - `macaroon`
+- Flip through each image and use Hydrus's "related tags" feature to quickly identify important tags that might be missing.
+  - This feature looks at other images with similar tags to provide suggestions.