Wan2.1 T2V 14B OldBookIllustrations

🚀 Wan2.1-T2V-14B Old Book Illustrations LoRA

This project is a LoRA adapter for the Wan-AI/Wan2.1-T2V-14B text - to - video model. It is trained on a subset of images from AdamLucek/oldbookillustrations-small, enabling the generation of video content in the style of old book illustrations.

🚀 Quick Start

Trigger words

You should use An old book illustration of a to trigger the image generation.

✨ Features

This LoRA adapter allows you to generate video content with the style of old book illustrations based on the Wan-AI/Wan2.1-T2V-14B text - to - video model.

📦 Installation

Using with Diffusers

pip install git+https://github.com/huggingface/diffusers.git

💻 Usage Examples

Using with Diffusers

import torch
from diffusers.utils import export_to_video
from diffusers import AutoencoderKLWan, WanPipeline
from diffusers.schedulers.scheduling_unipc_multistep import UniPCMultistepScheduler

model_id = "Wan-AI/Wan2.1-T2V-14B-Diffusers"
vae = AutoencoderKLWan.from_pretrained(model_id, subfolder="vae", torch_dtype=torch.float32)
pipe = WanPipeline.from_pretrained(model_id, vae=vae, torch_dtype=torch.bfloat16)
pipe.scheduler = UniPCMultistepScheduler.from_config(
  pipe.scheduler.config,
  flow_shift=5.0
)
pipe.to("cuda")
pipe.load_lora_weights("AdamLucek/Wan2.1-T2V-14B-OldBookIllustrations")
pipe.enable_model_cpu_offload() # for low-vram environments

prompt = "An old book illustration of a dog walking down a path"
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 = pipe(
    prompt=prompt,
    negative_prompt=negative_prompt,
    height=720,
    width=1280,
    num_frames=81,
    guidance_scale=5.0,
    num_inference_steps=32
).frames[0]
export_to_video(output, "output.mp4", fps=16)

Using with ComfyUI

Use the provided ComfyUI oldbookillustration_workflow.json.

To quickly download the recommended text encoder, VAE and Wan2.1 files run:

wget https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors
wget https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/vae/wan_2.1_vae.safetensors
wget https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/diffusion_models/wan2.1_t2v_14B_bf16.safetensors

📚 Documentation

Model Description

Lora adapter for Wan-AI/Wan2.1-T2V-14B text - 2 - video model trained on a subset of images from AdamLucek/oldbookillustrations-small.

📄 License

This project is licensed under the MIT license.

Download model

Weights for this model are available in Safetensors format.

Download them in the Files & versions tab.