README.md

---
license: apache-2.0
base_model: "Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
tags:
  - WanPipeline
  - WanPipeline-diffusers
  - text-to-image
  - image-to-image
  - diffusers
  - simpletuner
  - not-for-all-audiences
  - lora
  - template:sd-lora
  - standard
pipeline_tag: text-to-image
inference: true
widget:
- text: 'A black and white animated scene unfolds featuring a distressed upright cow with prominent horns and expressive eyes, suspended by its legs from a hook on a static background wall. A smaller Mickey Mouse-like character enters, standing near a wooden bench, initiating interaction between the two. The cow''s posture changes as it leans, stretches, and falls, while the mouse watches with a concerned expression, its face a mixture of curiosity and worry, in a world devoid of color.'
  parameters:
    negative_prompt: '色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走'
  output:
    url: ./assets/image_0_0.gif
---

# wan-disney-DCM-distilled

This is a standard PEFT LoRA derived from [Wan-AI/Wan2.1-T2V-1.3B-Diffusers](https://huggingface.co/Wan-AI/Wan2.1-T2V-1.3B-Diffusers).

The main validation prompt used during training was:
```
A black and white animated scene unfolds featuring a distressed upright cow with prominent horns and expressive eyes, suspended by its legs from a hook on a static background wall. A smaller Mickey Mouse-like character enters, standing near a wooden bench, initiating interaction between the two. The cow's posture changes as it leans, stretches, and falls, while the mouse watches with a concerned expression, its face a mixture of curiosity and worry, in a world devoid of color.
```


## Validation settings
- CFG: `1.0`
- CFG Rescale: `0.0`
- Steps: `8`
- Sampler: `FlowMatchEulerDiscreteScheduler`
- Seed: `42`
- Resolution: `832x480`


Note: The validation settings are not necessarily the same as the [training settings](#training-settings).

You can find some example images in the following gallery:


<Gallery />

The text encoder **was not** trained.
You may reuse the base model text encoder for inference.


## Training settings

- Training epochs: 0
- Training steps: 300
- Learning rate: 0.0001
  - Learning rate schedule: cosine
  - Warmup steps: 400000
- Max grad value: 0.01
- Effective batch size: 2
  - Micro-batch size: 2
  - Gradient accumulation steps: 1
  - Number of GPUs: 1
- Gradient checkpointing: True
- Prediction type: flow_matching (extra parameters=['shift=17.0'])
- Optimizer: adamw_bf16
- Trainable parameter precision: Pure BF16
- Base model precision: `int8-quanto`
- Caption dropout probability: 0.1%


- LoRA Rank: 128
- LoRA Alpha: 128.0
- LoRA Dropout: 0.1
- LoRA initialisation style: default
    

## Datasets

### disney-black-and-white-wan
- Repeats: 10
- Total number of images: 68
- Total number of aspect buckets: 1
- Resolution: 0.2304 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No


## Inference


```python
import torch
from diffusers import DiffusionPipeline

model_id = 'Wan-AI/Wan2.1-T2V-1.3B-Diffusers'
adapter_id = 'bghira/wan-disney-DCM-distilled'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "A black and white animated scene unfolds featuring a distressed upright cow with prominent horns and expressive eyes, suspended by its legs from a hook on a static background wall. A smaller Mickey Mouse-like character enters, standing near a wooden bench, initiating interaction between the two. The cow's posture changes as it leans, stretches, and falls, while the mouse watches with a concerned expression, its face a mixture of curiosity and worry, in a world devoid of color."
negative_prompt = '色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走'

## Optional: quantise the model to save on vram.
## Note: The model was quantised during training, and so it is recommended to do the same during inference time.
from optimum.quanto import quantize, freeze, qint8
quantize(pipeline.transformer, weights=qint8)
freeze(pipeline.transformer)
    
pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu') # the pipeline is already in its target precision level
model_output = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    num_inference_steps=8,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(42),
    width=832,
    height=480,
    guidance_scale=1.0,
).images[0]

from diffusers.utils.export_utils import export_to_gif
export_to_gif(model_output, "output.gif", fps=15)

```