%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
EN
Wan2.1 Fun 14B Control
A text-to-video model supporting multi-resolution training and first/last frame prediction
Downloads 10.53k
Release Time : 3/26/2025
Model Overview
Wan2.1-Fun-1.3B is a text-to-video generation model that supports multi-resolution training and first/last frame prediction, capable of producing high-quality video content based on text descriptions.
Model Features
Multi-resolution support
Supports video generation at various resolutions including 512/768/1024
First/last frame prediction
Predicts the first and last frames of videos to enhance generation coherence
Multilingual support
Supports text input in both Chinese and English
Model Capabilities
Text-to-video
Image-to-video
Video-to-video
Controlled video generation
Use Cases
Creative content generation
Short video creation
Automatically generates creative short videos based on text descriptions
Produces high-quality, coherent video content
Advertisement production
Quickly generates product showcase videos
Supports multiple resolutions and control conditions
Education and training
Educational video generation
Automatically generates demonstration videos based on teaching content
Supports teaching materials in both Chinese and English
๐ Wan-Fun
๐ Welcome! This project is designed for text-to-video generation, offering a powerful solution for creating high - quality videos from text inputs.
๐ Quick Start
1. Cloud Usage: AliyunDSW/Docker
a. Via Alibaba Cloud DSW
DSW offers free GPU time, which users can apply for once. The free GPU time is valid for 3 months after application.
Alibaba Cloud provides free GPU time on Freetier. You can obtain it and use it in Alibaba Cloud PAI - DSW. CogVideoX - Fun can be launched within 5 minutes.
b. Via ComfyUI
Our ComfyUI interface is as follows. For details, check ComfyUI README.
c. Via Docker
If you use Docker, make sure that the graphics card driver and CUDA environment are correctly installed on your machine. Then execute the following commands in sequence:
# pull image
docker pull mybigpai-public-registry.cn-beijing.cr.aliyuncs.com/easycv/torch_cuda:cogvideox_fun
# enter image
docker run -it -p 7860:7860 --network host --gpus all --security-opt seccomp:unconfined --shm-size 200g mybigpai-public-registry.cn-beijing.cr.aliyuncs.com/easycv/torch_cuda:cogvideox_fun
# clone code
git clone https://github.com/aigc-apps/CogVideoX-Fun.git
# enter CogVideoX-Fun's dir
cd CogVideoX-Fun
# download weights
mkdir models/Diffusion_Transformer
mkdir models/Personalized_Model
# Please use the hugginface link or modelscope link to download the model.
# CogVideoX-Fun
# https://huggingface.co/alibaba-pai/CogVideoX-Fun-V1.1-5b-InP
# https://modelscope.cn/models/PAI/CogVideoX-Fun-V1.1-5b-InP
# Wan
# https://huggingface.co/alibaba-pai/Wan2.1-Fun-14B-InP
# https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP
2. Local Installation: Environment Check/Download/Installation
a. Environment Check
We have verified that this library can be executed in the following environments:
Details for Windows:
- Operating System: Windows 10
- Python: python3.10 & python3.11
- PyTorch: torch2.2.0
- CUDA: 11.8 & 12.1
- CUDNN: 8+
- GPU: Nvidia - 3060 12G & Nvidia - 3090 24G
Details for Linux:
- Operating System: Ubuntu 20.04, CentOS
- Python: python3.10 & python3.11
- PyTorch: torch2.2.0
- CUDA: 11.8 & 12.1
- CUDNN: 8+
- GPU: Nvidia - V100 16G & Nvidia - A10 24G & Nvidia - A100 40G & Nvidia - A100 80G
You need approximately 60GB of available disk space. Please check!
b. Weight Placement
It is recommended to place the weights in the specified paths:
๐ฆ models/
โโโ ๐ Diffusion_Transformer/
โ โโโ ๐ CogVideoX-Fun-V1.1-2b-InP/
โ โโโ ๐ CogVideoX-Fun-V1.1-5b-InP/
โ โโโ ๐ Wan2.1-Fun-14B-InP
โ โโโ ๐ Wan2.1-Fun-1.3B-InP/
โโโ ๐ Personalized_Model/
โ โโโ your trained trainformer model / your trained lora model (for UI load)
โจ Features
Model Address
V1.0:
| Name | Storage Space | Hugging Face | Model Scope | Description |
|---|---|---|---|---|
| Wan2.1 - Fun - 1.3B - InP | 19.0 GB | ๐คLink | ๐Link | The weights for text - to - video generation of Wan2.1 - Fun - 1.3B, trained at multiple resolutions, support the prediction of the first and last frames. |
| Wan2.1 - Fun - 14B - InP | 47.0 GB | ๐คLink | ๐Link | The weights for text - to - video generation of Wan2.1 - Fun - 14B, trained at multiple resolutions, support the prediction of the first and last frames. |
| Wan2.1 - Fun - 1.3B - Control | 19.0 GB | ๐คLink | ๐Link | The video control weights of Wan2.1 - Fun - 1.3B, support different control conditions such as Canny, Depth, Pose, MLSD, etc., and also support trajectory control. Support video prediction at multiple resolutions (512, 768, 1024), trained with 81 frames at 16 frames per second, support multi - language prediction. |
| Wan2.1 - Fun - 14B - Control | 47.0 GB | ๐คLink | ๐Link | The video control weights of Wan2.1 - Fun - 14B, support different control conditions such as Canny, Depth, Pose, MLSD, etc., and also support trajectory control. Support video prediction at multiple resolutions (512, 768, 1024), trained with 81 frames at 16 frames per second, support multi - language prediction. |
Video Works
Wan2.1 - Fun - 14B - InP && Wan2.1 - Fun - 1.3B - InP
Wan2.1 - Fun - 14B - Control && Wan2.1 - Fun - 1.3B - Control
๐ป Usage Examples
1. Generation
a. Memory - Saving Solution
Since the parameters of Wan2.1 are very large, we need to consider a memory - saving solution to adapt to consumer - grade graphics cards. We provide a GPU_memory_mode for each prediction file, which can be selected from model_cpu_offload, model_cpu_offload_and_qfloat8, and sequential_cpu_offload. This solution is also applicable to the generation of CogVideoX - Fun.
model_cpu_offloadmeans that the entire model will be transferred to the CPU after use, which can save some video memory.model_cpu_offload_and_qfloat8means that the entire model will be transferred to the CPU after use, and the transformer model is quantized to float8, which can save more video memory.sequential_cpu_offloadmeans that each layer of the model will be transferred to the CPU after use. It is slower but can save a large amount of video memory.
qfloat8 will partially reduce the performance of the model but can save more video memory. If the video memory is sufficient, it is recommended to use model_cpu_offload.
b. Via ComfyUI
For details, check ComfyUI README.
c. Run Python Files
- Step 1: Download the corresponding weights and place them in the
modelsfolder. - Step 2: Use different files for prediction according to different weights and prediction targets. Currently, this library supports CogVideoX - Fun, Wan2.1, and Wan2.1 - Fun, which are distinguished by folder names in the
examplesfolder. Different models support different functions, so please make distinctions according to the specific situation. Take CogVideoX - Fun as an example.- Text - to - video:
- Modify
prompt,neg_prompt,guidance_scale, andseedin theexamples/cogvideox_fun/predict_t2v.pyfile. - Then run the
examples/cogvideox_fun/predict_t2v.pyfile and wait for the generation result. The result will be saved in thesamples/cogvideox - fun - videosfolder.
- Modify
- Image - to - video:
- Modify
validation_image_start,validation_image_end,prompt,neg_prompt,guidance_scale, andseedin theexamples/cogvideox_fun/predict_i2v.pyfile. validation_image_startis the starting image of the video, andvalidation_image_endis the ending image of the video.- Then run the
examples/cogvideox_fun/predict_i2v.pyfile and wait for the generation result. The result will be saved in thesamples/cogvideox - fun - videos_i2vfolder.
- Modify
- Video - to - video:
- Modify
validation_video,validation_image_end,prompt,neg_prompt,guidance_scale, andseedin theexamples/cogvideox_fun/predict_v2v.pyfile. validation_videois the reference video for video - to - video generation. You can use the following video for demonstration: Demo Video- Then run the
examples/cogvideox_fun/predict_v2v.pyfile and wait for the generation result. The result will be saved in thesamples/cogvideox - fun - videos_v2vfolder.
- Modify
- Normal control - to - video (Canny, Pose, Depth, etc.):
- Modify
control_video,validation_image_end,prompt,neg_prompt,guidance_scale, andseedin theexamples/cogvideox_fun/predict_v2v_control.pyfile. control_videois the control video for control - to - video generation, which is a video extracted by operators such as Canny, Pose, Depth, etc. You can use the following video for demonstration: Demo Video- Then run the
examples/cogvideox_fun/predict_v2v_control.pyfile and wait for the generation result. The result will be saved in thesamples/cogvideox - fun - videos_v2v_controlfolder.
- Modify
- Text - to - video:
- Step 3: If you want to combine other backbones and LoRA models you trained, modify
lora_pathinexamples/{model_name}/predict_t2v.pyandexamples/{model_name}/predict_i2v.pyaccording to the situation.
d. Via UI Interface
The webui supports text - to - video, image - to - video, video - to - video, and normal control - to - video (Canny, Pose, Depth, etc.). Currently, this library supports CogVideoX - Fun, Wan2.1, and Wan2.1 - Fun, which are distinguished by folder names in the examples folder. Different models support different functions, so please make distinctions according to the specific situation. Take CogVideoX - Fun as an example.
- Step 1: Download the corresponding weights and place them in the
modelsfolder. - Step 2: Run the
examples/cogvideox_fun/app.pyfile and enter the Gradio page. - Step 3: Select the generation model on the page, fill in
prompt,neg_prompt,guidance_scale,seed, etc., click "Generate", and wait for the generation result. The result will be saved in thesamplefolder.
๐ Documentation
References
- CogVideo: https://github.com/THUDM/CogVideo/
- EasyAnimate: https://github.com/aigc-apps/EasyAnimate
- Wan2.1: https://github.com/Wan-Video/Wan2.1/
๐ License
This project is licensed under the Apache License (Version 2.0).
Xclip Base Patch32
MIT
X-CLIP is an extended version of CLIP for general video-language understanding, trained on (video, text) pairs via contrastive learning, suitable for tasks like video classification and video-text retrieval.
Text-to-Video
Transformers English
Transformers EnglishX
microsoft
309.80k
84
LTX Video
Other
The first DiT-based video generation model capable of real-time generation of high-quality videos, supporting two scenarios: text-to-video and image + text-to-video.
Text-to-Video English
L
Lightricks
165.42k
1,174
Wan2.1 14B VACE GGUF
Apache-2.0
The GGUF format version of the Wan2.1-VACE-14B model, mainly used for text-to-video generation tasks.
Text-to-Video
W
QuantStack
146.36k
139
Animatediff Lightning
Openrail
Ultra-fast text-to-video model, generating videos over ten times faster than the original AnimateDiff
Text-to-Video
A
ByteDance
144.00k
925
V Express
V-Express is an audio and facial keypoint condition-based video generation model capable of converting audio input into dynamic video output.
Text-to-Video English
V
tk93
118.36k
85
Cogvideox 5b
Other
CogVideoX is the open-source version of the video generation model derived from Qingying, providing high-quality video generation capabilities.
Text-to-Video English
C
THUDM
92.32k
611
Llava NeXT Video 7B Hf
LLaVA-NeXT-Video is an open-source multimodal chatbot that achieves excellent video understanding capabilities through mixed training on video and image data, reaching SOTA level among open-source models on the VideoMME benchmark.
Text-to-Video Transformers English
Transformers EnglishL
llava-hf
65.95k
88
Wan2.1 T2V 14B Diffusers
Apache-2.0
Wan2.1 is a comprehensive open-source video foundation model designed to push the boundaries of video generation, supporting tasks such as text-to-video in Chinese and English, image-to-video, and more.
Text-to-Video Supports Multiple Languages
W
Wan-AI
48.65k
24
Wan2.1 T2V 1.3B Diffusers
Apache-2.0
Wan 2.1 is a comprehensive open-source video foundation model featuring top-tier performance, consumer-grade GPU support, multi-task capabilities, visual-text generation, and efficient video VAE.
Text-to-Video Supports Multiple Languages
W
Wan-AI
45.29k
38
Wan2.1 T2V 14B
Apache-2.0
Wan 2.1 is a comprehensive open-source video foundation model capable of multiple tasks including text-to-video, image-to-video, video editing, text-to-image, and video-to-audio generation, with support for bilingual Chinese-English text input.
Text-to-Video Supports Multiple Languages
W
Wan-AI
44.88k
1,238
Featured Recommended AI Models
ยฉ 2025AIbase