Experience the CogVideoX-5B model online at π€ Huggingface Space or π€ ModelScope Space
π View the paper and user guide
π Join our WeChat and Discord
π Visit QingYing and API Platform to experience larger-scale commercial video generation models.
- π₯π₯ News:
2024/10/13: A more cost-effective fine-tuning framework forCogVideoX-5Bthat works with a single 4090 GPU, cogvideox-factory, has been released. It supports fine-tuning with multiple resolutions. Feel free to use it! - π₯ News:
2024/10/10: We have updated our technical report. Please click here to view it. More training details and a demo have been added. To see the demo, click here.- π₯ News:2024/10/09: We have publicly released the technical documentation for CogVideoX fine-tuning on Feishu, further increasing distribution flexibility. All examples in the public documentation can be fully reproduced. - π₯ News:
2024/9/19: We have open-sourced the CogVideoX series image-to-video model CogVideoX-5B-I2V. This model can take an image as a background input and generate a video combined with prompt words, offering greater controllability. With this, the CogVideoX series models now support three tasks: text-to-video generation, video continuation, and image-to-video generation. Welcome to try it online at Experience. - π₯
2024/9/19: The Caption model CogVLM2-Caption, used in the training process of CogVideoX to convert video data into text descriptions, has been open-sourced. Welcome to download and use it. - π₯
2024/8/27: We have open-sourced a larger model in the CogVideoX series, CogVideoX-5B. We have significantly optimized the model's inference performance, greatly lowering the inference threshold. You can run * CogVideoX-2B* on older GPUs likeGTX 1080TI, and CogVideoX-5B on desktop GPUs likeRTX 3060. Please strictly follow the requirements to update and install dependencies, and refer to cli_demo for inference code. Additionally, the open-source license for the **CogVideoX-2B ** model has been changed to the Apache 2.0 License. - π₯
2024/8/6: We have open-sourced 3D Causal VAE, used for CogVideoX-2B, which can reconstruct videos with almost no loss. - π₯
2024/8/6: We have open-sourced the first model of the CogVideoX series video generation models, **CogVideoX-2B **. - π± Source:
2022/5/19: We have open-sourced the CogVideo video generation model (now you can see it in theCogVideobranch). This is the first open-source large Transformer-based text-to-video generation model. You can access the ICLR'23 paper for technical details.
Jump to a specific section:
- Quick Start
- CogVideoX-2B Video Works
- Introduction to the CogVideoX Model
- Full Project Structure
- Introduction to CogVideo(ICLR'23) Model
- Citations
- Open Source Project Plan
- Model License
Before running the model, please refer to this guide to see how we use large models like GLM-4 (or other comparable products, such as GPT-4) to optimize the model. This is crucial because the model is trained with long prompts, and a good prompt directly impacts the quality of the video generation.
Please make sure your Python version is between 3.10 and 3.12, inclusive of both 3.10 and 3.12.
Follow instructions in sat_demo: Contains the inference code and fine-tuning code of SAT weights. It is recommended to improve based on the CogVideoX model structure. Innovative researchers use this code to better perform rapid stacking and development.
Please make sure your Python version is between 3.10 and 3.12, inclusive of both 3.10 and 3.12.
pip install -r requirements.txt
Then follow diffusers_demo: A more detailed explanation of the inference code, mentioning the significance of common parameters.
For more details on quantized inference, please refer to diffusers-torchao. With Diffusers and TorchAO, quantized inference is also possible leading to memory-efficient inference as well as speedup in some cases when compiled. A full list of memory and time benchmarks with various settings on A100 and H100 has been published at diffusers-torchao.
5b_1.mp4 |
5b_2.mp4 |
5b_3.mp4 |
5b_4.mp4 |
5b_5.mp4 |
5b_6.mp4 |
5b_7.mp4 |
5b_8.mp4 |
1.mp4 |
2.mp4 |
3.mp4 |
4.mp4 |
To view the corresponding prompt words for the gallery, please click here
CogVideoX is an open-source version of the video generation model originating from QingYing. The table below displays the list of video generation models we currently offer, along with their foundational information.
| Model Name | CogVideoX-2B | CogVideoX-5B | CogVideoX-5B-I2V |
|---|---|---|---|
| Model Description | Entry-level model, balancing compatibility. Low cost for running and secondary development. | Larger model with higher video generation quality and better visual effects. | CogVideoX-5B image-to-video version. |
| Inference Precision | FP16*(recommended), BF16, FP32, FP8*, INT8, not supported: INT4 | BF16 (recommended), FP16, FP32, FP8*, INT8, not supported: INT4 | |
| Single GPU Memory Usage |
SAT FP16: 18GB diffusers FP16: from 4GB* diffusers INT8 (torchao): from 3.6GB* |
SAT BF16: 26GB diffusers BF16: from 5GB* diffusers INT8 (torchao): from 4.4GB* |
|
| Multi-GPU Inference Memory Usage | FP16: 10GB* using diffusers |
BF16: 15GB* using diffusers |
|
| Inference Speed (Step = 50, FP/BF16) |
Single A100: ~90 seconds Single H100: ~45 seconds |
Single A100: ~180 seconds Single H100: ~90 seconds |
|
| Fine-tuning Precision | FP16 | BF16 | |
| Fine-tuning Memory Usage | 47 GB (bs=1, LORA) 61 GB (bs=2, LORA) 62GB (bs=1, SFT) |
63 GB (bs=1, LORA) 80 GB (bs=2, LORA) 75GB (bs=1, SFT) |
78 GB (bs=1, LORA) 75GB (bs=1, SFT, 16GPU) |
| Prompt Language | English* | ||
| Maximum Prompt Length | 226 Tokens | ||
| Video Length | 6 Seconds | ||
| Frame Rate | 8 Frames / Second | ||
| Video Resolution | 720 x 480, no support for other resolutions (including fine-tuning) | ||
| Position Encoding | 3d_sincos_pos_embed | 3d_sincos_pos_embed | 3d_rope_pos_embed + learnable_pos_embed |
| Download Link (Diffusers) | π€ HuggingFace π€ ModelScope π£ WiseModel |
π€ HuggingFace π€ ModelScope π£ WiseModel |
π€ HuggingFace π€ ModelScope π£ WiseModel |
| Download Link (SAT) | SAT | ||
Data Explanation
- While testing using the diffusers library, all optimizations included in the diffusers library were enabled. This scheme has not been tested for actual memory usage on devices outside of NVIDIA A100 / H100 architectures. Generally, this scheme can be adapted to all NVIDIA Ampere architecture and above devices. If optimizations are disabled, memory consumption will multiply, with peak memory usage being about 3 times the value in the table. However, speed will increase by about 3-4 times. You can selectively disable some optimizations, including:
pipe.enable_sequential_cpu_offload()
pipe.vae.enable_slicing()
pipe.vae.enable_tiling()
- For multi-GPU inference, the
enable_sequential_cpu_offload()optimization needs to be disabled. - Using INT8 models will slow down inference, which is done to accommodate lower-memory GPUs while maintaining minimal video quality loss, though inference speed will significantly decrease.
- The CogVideoX-2B model was trained in
FP16precision, and all CogVideoX-5B models were trained inBF16precision. We recommend using the precision in which the model was trained for inference. - PytorchAO and Optimum-quanto can be
used to quantize the text encoder, transformer, and VAE modules to reduce the memory requirements of CogVideoX. This
allows the model to run on free T4 Colabs or GPUs with smaller memory! Also, note that TorchAO quantization is fully
compatible with
torch.compile, which can significantly improve inference speed. FP8 precision must be used on devices with NVIDIA H100 and above, requiring source installation oftorch,torchao,diffusers, andacceleratePython packages. CUDA 12.4 is recommended. - The inference speed tests also used the above memory optimization scheme. Without memory optimization, inference speed
increases by about 10%. Only the
diffusersversion of the model supports quantization. - The model only supports English input; other languages can be translated into English for use via large model refinement.
- The memory usage of model fine-tuning is tested in an
8 * H100environment, and the program automatically usesZero 2optimization. If a specific number of GPUs is marked in the table, that number or more GPUs must be used for fine-tuning.
We highly welcome contributions from the community and actively contribute to the open-source community. The following works have already been adapted for CogVideoX, and we invite everyone to use them:
- CogVideoX-Fun: CogVideoX-Fun is a modified pipeline based on the CogVideoX architecture, supporting flexible resolutions and multiple launch methods.
- CogStudio: A separate repository for CogVideo's Gradio Web UI, which supports more functional Web UIs.
- Xorbits Inference: A powerful and comprehensive distributed inference framework, allowing you to easily deploy your own models or the latest cutting-edge open-source models with just one click.
- ComfyUI-CogVideoXWrapper Use the ComfyUI framework to integrate CogVideoX into your workflow.
- VideoSys: VideoSys provides a user-friendly, high-performance infrastructure for video generation, with full pipeline support and continuous integration of the latest models and techniques.
- AutoDL Space: A one-click deployment Huggingface Space image provided by community members.
- Interior Design Fine-Tuning Model: is a fine-tuned model based on CogVideoX, specifically designed for interior design.
- xDiT: xDiT is a scalable inference engine for Diffusion Transformers (DiTs)
on multiple GPU Clusters. xDiT supports real-time image and video generations services.
cogvideox-factory: A cost-effective
fine-tuning framework for CogVideoX, compatible with the
diffusersversion model. Supports more resolutions, and fine-tuning CogVideoX-5B can be done with a single 4090 GPU. - CogVideoX-Interpolation: A pipeline based on the modified CogVideoX structure, aimed at providing greater flexibility for keyframe interpolation generation.
- DiffSynth-Studio: DiffSynth Studio is a diffusion engine. It has restructured the architecture, including text encoders, UNet, VAE, etc., enhancing computational performance while maintaining compatibility with open-source community models. The framework has been adapted for CogVideoX.
This open-source repository will guide developers to quickly get started with the basic usage and fine-tuning examples of the CogVideoX open-source model.
Here provide three projects that can be run directly on free Colab T4 instances:
- CogVideoX-5B-T2V-Colab.ipynb: CogVideoX-5B Text-to-Video Colab code.
- CogVideoX-5B-T2V-Int8-Colab.ipynb: CogVideoX-5B Quantized Text-to-Video Inference Colab code, which takes about 30 minutes per run.
- CogVideoX-5B-I2V-Colab.ipynb: CogVideoX-5B Image-to-Video Colab code.
- CogVideoX-5B-V2V-Colab.ipynb: CogVideoX-5B Video-to-Video Colab code.
- dcli_demo: A more detailed inference code explanation, including the significance of common parameters. All of this is covered here.
- cli_demo_quantization: Quantized model inference code that can run on devices with lower memory. You can also modify this code to support running CogVideoX models in FP8 precision.
- diffusers_vae_demo: Code for running VAE inference separately.
- space demo: The same GUI code as used in the Huggingface Space, with frame interpolation and super-resolution tools integrated.
- convert_demo: How to convert user input into long-form input suitable for CogVideoX. Since CogVideoX is trained on long texts, we need to transform the input text distribution to match the training data using an LLM. The script defaults to using GLM-4, but it can be replaced with GPT, Gemini, or any other large language model.
- gradio_web_demo: A simple Gradio web application demonstrating how to use the CogVideoX-2B / 5B model to generate videos. Similar to our Huggingface Space, you can use this script to run a simple web application for video generation.
- finetune_demo: Fine-tuning scheme and details of the diffusers version of the CogVideoX model.
- sat_demo: Contains the inference code and fine-tuning code of SAT weights. It is recommended to improve based on the CogVideoX model structure. Innovative researchers use this code to better perform rapid stacking and development.
This folder contains some tools for model conversion / caption generation, etc.
- convert_weight_sat2hf: Converts SAT model weights to Huggingface model weights.
- caption_demo: Caption tool, a model that understands videos and outputs descriptions in text.
- export_sat_lora_weight: SAT fine-tuning model export tool, exports the SAT Lora Adapter in diffusers format.
- load_cogvideox_lora: Tool code for loading the diffusers version of fine-tuned Lora Adapter.
- llm_flux_cogvideox: Automatically generate videos using an open-source local large language model + Flux + CogVideoX.
- parallel_inference_xdit: Supported by xDiT, parallelize the video generation process on multiple GPUs.
The official repo for the paper: CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformers is on the CogVideo branch
CogVideo is able to generate relatively high-frame-rate videos. A 4-second clip of 32 frames is shown below.
cogvideo.mp4
The demo for CogVideo is at https://models.aminer.cn/cogvideo, where you can get hands-on practice on text-to-video generation. The original input is in Chinese.
π If you find our work helpful, please leave us a star and cite our paper.
@article{yang2024cogvideox,
title={CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer},
author={Yang, Zhuoyi and Teng, Jiayan and Zheng, Wendi and Ding, Ming and Huang, Shiyu and Xu, Jiazheng and Yang, Yuanming and Hong, Wenyi and Zhang, Xiaohan and Feng, Guanyu and others},
journal={arXiv preprint arXiv:2408.06072},
year={2024}
}
@article{hong2022cogvideo,
title={CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformers},
author={Hong, Wenyi and Ding, Ming and Zheng, Wendi and Liu, Xinghan and Tang, Jie},
journal={arXiv preprint arXiv:2205.15868},
year={2022}
}
We welcome your contributions! You can click here for more information.
The code in this repository is released under the Apache 2.0 License.
The CogVideoX-2B model (including its corresponding Transformers module and VAE module) is released under the Apache 2.0 License.
The CogVideoX-5B model (Transformers module, include I2V and T2V) is released under the CogVideoX LICENSE.