|
--- |
|
license: bsd-3-clause |
|
language: |
|
- en |
|
- zh |
|
pipeline_tag: video-text-to-text |
|
--- |
|
|
|
# Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding |
|
This is the Hugging Face repo for storing pre-trained & fine-tuned checkpoints of our [Video-LLaMA](https://arxiv.org/abs/2306.02858), which is a multi-modal conversational large language model with video understanding capability. |
|
|
|
|
|
## Vision-Language Branch |
|
| Checkpoint | Link | Note | |
|
|:------------|-------------|-------------| |
|
| pretrain-vicuna7b | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/pretrain_vicuna7b-v2.pth) | Pre-trained on WebVid (2.5M video-caption pairs) and LLaVA-CC3M (595k image-caption pairs) | |
|
| finetune-vicuna7b-v2 | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/finetune-vicuna7b-v2.pth) | Fine-tuned on the instruction-tuning data from [MiniGPT-4](https://github.com/Vision-CAIR/MiniGPT-4), [LLaVA](https://github.com/haotian-liu/LLaVA) and [VideoChat](https://github.com/OpenGVLab/Ask-Anything)| |
|
| pretrain-vicuna13b | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/pretrain-vicuna13b.pth) | Pre-trained on WebVid (2.5M video-caption pairs) and LLaVA-CC3M (595k image-caption pairs) | |
|
| finetune-vicuna13b-v2 | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/finetune-vicuna13b-v2.pth) | Fine-tuned on the instruction-tuning data from [MiniGPT-4](https://github.com/Vision-CAIR/MiniGPT-4), [LLaVA](https://github.com/haotian-liu/LLaVA) and [VideoChat](https://github.com/OpenGVLab/Ask-Anything)| |
|
| pretrain-ziya13b-zh | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/pretrain-ziya13b-zh.pth) | Pre-trained with Chinese LLM [Ziya-13B](https://huggingface.co/IDEA-CCNL/Ziya-LLaMA-13B-v1) | |
|
| finetune-ziya13b-zh | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/finetune-ziya13b-zh.pth) | Fine-tuned on machine-translated [VideoChat](https://github.com/OpenGVLab/Ask-Anything) instruction-following dataset (in Chinese)| |
|
| pretrain-billa7b-zh | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/pretrain-billa7b-zh.pth) | Pre-trained with Chinese LLM [BiLLA-7B](https://huggingface.co/IDEA-CCNL/Ziya-LLaMA-13B-v1) | |
|
| finetune-billa7b-zh | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/finetune-billa7b-zh.pth) | Fine-tuned on machine-translated [VideoChat](https://github.com/OpenGVLab/Ask-Anything) instruction-following dataset (in Chinese) | |
|
|
|
## Audio-Language Branch |
|
| Checkpoint | Link | Note | |
|
|:------------|-------------|-------------| |
|
| pretrain-vicuna7b | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/pretrain_vicuna7b_audiobranch.pth) | Pre-trained on WebVid (2.5M video-caption pairs) and LLaVA-CC3M (595k image-caption pairs) | |
|
| finetune-vicuna7b-v2 | [link](https://huggingface.co/DAMO-NLP-SG/Video-LLaMA-Series/resolve/main/finetune_vicuna7b_audiobranch.pth) | Fine-tuned on the instruction-tuning data from [MiniGPT-4](https://github.com/Vision-CAIR/MiniGPT-4), [LLaVA](https://github.com/haotian-liu/LLaVA) and [VideoChat](https://github.com/OpenGVLab/Ask-Anything)| |
|
|
|
|
|
## Usage |
|
For launching the pre-trained Video-LLaMA on your own machine, please refer to our [github repo](https://github.com/DAMO-NLP-SG/Video-LLaMA). |
