---
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- text: >-
    1girl, solo, blue background, halftone background, looking at viewer, animal
    ears, school uniform, yellow eyes, black hair, long hair, sitting, crossed
    legs, cat ears, border, halftone, white border, couch, open mouth, shirt,
    skirt, long sleeves, white shirt, black skirt, pleated skirt, serafuku,
    neckerchief, black sailor collar, socks, white socks, outside border, sailor
    collar, animal ear fluff, black neckerchief, masterpiece, best quality, high
    score, great score, latest, figma
  parameters:
    negative_prompt: >-
      lowres, bad anatomy, bad hands, text, error, missing finger, extra digits,
      fewer digits, cropped, worst quality, low quality, low score, bad score,
      average score, signature, watermark, username, blurry, 
  output:
    url: images/image.jpg
- text: >-
    pvc, 1girl, solo, full body, gloves, groin, hair between eyes, hair
    ornament, head tilt, holding, holding staff, horns, long hair, looking at
    viewer, maid headdress, mole, mole under eye, navel, parted lips, purple
    eyes, purple hair, purple theme, shawl, shorts, single horn, staff,
    standing, thighhighs, twintails, white footwear, white gloves, white
    thighhighs, masterpiece, best quality, high score, great score, latest
  parameters:
    negative_prompt: >-
      lowres, bad anatomy, bad hands, text, error, missing finger, extra digits,
      fewer digits, cropped, worst quality, low quality, low score, bad score,
      average score, signature, watermark, username, blurry, 
  output:
    url: images/image(1).jpg
- text: >-
    cute anime girl with massive fluffy fennec ears and a big fluffy tail blonde
    messy long hair blue eyes wearing a maid outfit with a long black gold leaf
    pattern dress and a white apron mouth open placing a fancy black forest cake
    with candles on top of a dinner table of an old dark Victorian mansion lit
    by candlelight with a bright window to the foggy forest and very expensive
    stuff everywhere there are paintings on the walls, pvc, masterpiece, best
    quality, high score, great score, latest
  parameters:
    negative_prompt: >-
      lowres, bad anatomy, bad hands, text, error, missing finger, extra digits,
      fewer digits, cropped, worst quality, low quality, low score, bad score,
      average score, signature, watermark, username, blurry, 
  output:
    url: images/image(2).jpg
- text: >-
    1girl, solo, outdoors, looking at viewer, flower, gloves, grey hair, hat,
    jacket, long hair, long skirt, long sleeves, looking at viewer, open
    clothes, open jacket, pantyhose, red eyes, red flower, red hat, red jacket,
    red rose, red skirt, rose, shirt, skirt, smile, snowing, standing, white
    gloves, white shirt, pvc, masterpiece, best quality, high score, great
    score, latest
  parameters:
    negative_prompt: >-
      lowres, bad anatomy, bad hands, text, error, missing finger, extra digits,
      fewer digits, cropped, worst quality, low quality, low score, bad score,
      average score, signature, watermark, username, blurry, 
  output:
    url: images/image(3).jpg
- text: >-
    1girl, solo, maid, cowboy shot, cup, green eyes, green hair, hair intakes,
    hair ornament, holding, holding cup, lips, long hair, looking at viewer,
    ponytail, realistic, simple background, teacup, very long hair, white
    background, wrist cuffs, pvc, masterpiece, best quality, high score, great
    score, latest
  parameters:
    negative_prompt: >-
      lowres, bad anatomy, bad hands, text, error, missing finger, extra digits,
      fewer digits, cropped, worst quality, low quality, low score, bad score,
      average score, signature, watermark, username, blurry, 
  output:
    url: images/image(4).jpg
- text: >-
    1girl, hatsune miku, vocaloid, solo, :o, bare shoulders, black skirt, black
    sleeves, black thighhighs, blue eyes, blue hair, blush, collared shirt,
    detached sleeves, hair ornament, kneeling, long hair, long sleeves, looking
    at viewer, miniskirt, parted lips, pleated skirt, shirt, skirt, sleeveless,
    sleeveless shirt, thighhighs, twintails, very long hair, white shirt, wing
    collar, pvc, masterpiece, best quality, high score, great score, latest
  parameters:
    negative_prompt: >-
      lowres, bad anatomy, bad hands, text, error, missing finger, extra digits,
      fewer digits, cropped, worst quality, low quality, low score, bad score,
      average score, signature, watermark, username, blurry, 
  output:
    url: images/image(5).jpg
- text: >-
    1girl, solo, arm at side, black hairband, black shirt, black skirt, blue
    eyes, blush, collared shirt, dress shirt, gradient hair, hair between eyes,
    hairband, hand up, long hair, long sleeves, multicolored hair, neckerchief,
    orange hair, orange neckerchief, parted lips, pleated skirt, purple hair,
    shirt, sidelocks, simple background, skirt, sleeves past wrists, upper body,
    very long hair, white background, pvc, masterpiece, best quality, high
    score, great score, latest
  parameters:
    negative_prompt: >-
      lowres, bad anatomy, bad hands, text, error, missing finger, extra digits,
      fewer digits, cropped, worst quality, low quality, low score, bad score,
      average score, signature, watermark, username, blurry, 
  output:
    url: images/image(6).jpg
base_model: cagliostrolab/animagine-xl-4.0
instance_prompt: pvc, nendoroid, figma
license: apache-2.0
---
# PVC Style LoRA for AnimagineXL 4.0 opt

<Gallery />

## Model description 

Only tested with ComfyUI.


## Trigger words

You should use `pvc` to trigger the image generation.

You should use `nendoroid` to trigger the image generation.

You should use `figma` to trigger the image generation.


## Download model

Weights for this model are available in Safetensors format.

[Download](/p1atdev/animagine-xl-4.0-opt-pvc-style-lora/tree/main) them in the Files & versions tab.

## Dataset

About 7000 of pvc figure images.

## Training

### Training code

https://github.com/p1atdev/vision-ft

### Training settings


<details>
<summary>`config.yml`</summary>
  
```yml
model:
  checkpoint_path: "./models/animagine-xl-4.0-opt.bnb_nf4.safetensors"
  pretrained_model_name_or_path: cagliostrolab/animagine-xl-4.0

  dtype: bfloat16

  denoiser:
    attn_implementation: "flash_attention_2"

peft:
  config:
    type: lora
    rank: 2
    alpha: 1.0
    dropout: 0.0

    dtype: bfloat16

  include_keys:
    - "attn1"
    - "attn2"
    - ".ff."
  exclude_keys: ["text_encoder", "vae"]

dataset:
  folder: "data/pvc"
  num_repeats: 4
  batch_size: 4

  bucket_base_size: 1024
  step: 128
  min_size: 384
  do_upscale: true

  caption_processors:
    - type: shuffle
      split_separator: ","

optimizer:
  name: "schedulefree.RAdamScheduleFree"
  args:
    lr: 0.005

scheduler:
  # name: "torch.optim.lr_scheduler.ConstantLR"
  # args: {}

tracker:
  project_name: "sdxl-pvc-1"
  loggers:
    - wandb

saving:
  strategy:
    per_epochs: 0.25
    per_steps: null
    save_last: true

  callbacks:
    - type: "safentensors"
      name: "sdxl-pvc"
      save_dir: "./output/sdxl-pvc-2"

preview:
  strategy:
    per_epochs: 1
    per_steps: 100

  callbacks:
    - type: "local"
      save_dir: "./output/sdxl-lora/preview"

  data:
    path: "./projects/pvc-sdxl/preview.yml"

seed: 42
num_train_epochs: 10

trainer:
  # debug_mode: "1step"

  gradient_checkpointing: true

  torch_compile: true
  torch_compile_args:
    backend: eager
    mode: default
    fullgraph: false

  fp32_matmul_precision: "high"
  allow_tf32: true
```

</details>

<details>
<summary>`preview.yml` </summary>
  
```yml
- prompt: |-
    pvc style, 1girl, aqua eyes, baseball cap, blonde hair, closed mouth, earrings, 
    green background, hat, hoop earrings, jewelry, looking at viewer, 
    shirt, short hair, simple background, solo, upper body, yellow shirt, 
    masterpiece, high score, great score, absurdres
  negative_prompt: |-
    lowres, bad anatomy, bad hands, text, error, missing finger, extra digits, 
    fewer digits, cropped, worst quality, low quality, low score, bad score, 
    average score, signature, watermark, username, blurry
  height: 1024
  width: 1024
  cfg_scale: 5.0
  num_steps: 20

  seed: 0
```

</details>

The training was stopped in the 9th epoch, because of my room was too hot due to this training.

### Device and Hours

RTX 4070 Ti Super (VRAM 16GB) for 27 hours.