Skip to content

Latest commit

 

History

History
77 lines (58 loc) · 3.66 KB

LLaVA-NeXT-Video.md

File metadata and controls

77 lines (58 loc) · 3.66 KB
Raw

LLaVA-NeXT: A Strong Zero-shot Video Understanding Model

Contents

Demo

make sure you installed the LLaVA-NeXT model files via outside REAME.md

  1. Example model: lmms-lab/LLaVA-NeXT-Video-7B-DPO

  2. Prompt mode: vicuna_v1 (use mistral_direct for lmms-lab/LLaVA-NeXT-Video-34B-DPO)

  3. Sampled frames: 32 (Defines how many frames to sample from the video.)

  4. Spatial pooling stride: 2 (With original tokens for one frame at 24x24, if stride=2, then the tokens for one frame are 12x12.)

  5. Local video path: ./data/llava_video/video-chatgpt/evaluation/Test_Videos/v_Lf_7RurLgp0.mp4

To run a demo, execute:

bash scripts/video/demo/video_demo.sh ${Example model} ${Prompt mode} ${Sampled frames} ${Spatial pooling stride} True ${Video path at local}

Example:

bash scripts/video/demo/video_demo.sh lmms-lab/LLaVA-NeXT-Video-7B-DPO vicuna_v1 32 2 True ./data/llava_video/video-chatgpt/evaluation/Test_Videos/v_Lf_7RurLgp0.mp4

Evaluation

Preparation

Please download the evaluation data and its metadata from the following links:

  1. video-chatgpt: here.
  2. video_detail_description: here.
  3. activity_qa: here and here.

Organize the downloaded data into the following structure:

LLaVA-NeXT
├── llava
├── scripts
└── data
    └── llava_video
        ├── video-chatgpt
        │   ├── Test_Videos
        │   ├── consistency_qa.json
        │   ├── consistency_qa_test.json
        │   ├── consistency_qa_train.json
        ├── video_detail_description
        │   └── Test_Human_Annotated_Captions
        └── ActivityNet-QA
            ├── all_test
            ├── test_a.json
            └── test_b.json

Inference and Evaluation

Example for video detail description evaluation (additional scripts are available in scripts/eval):

bash scripts/video/eval/video_detail_description_eval_shard.sh ${Example model} ${Prompt mode} ${Sampled frames} ${Spatial pooling stride} True 8

Example:

bash scripts/eval/video_detail_description_eval_shard.sh liuhaotian/llava-v1.6-vicuna-7b vicuna_v1 32 2 True 8

GPT Evaluation Example (Optional if the above step is completed)

Assuming you have pred.json (model-generated predictions) for model llava-v1.6-vicuna-7b at ./work_dirs/eval_video_detail_description/llava-v1.6-vicuna-7b_vicuna_v1_frames_32_stride_2:

bash scripts/video/eval/video_description_eval_only.sh llava-v1.6-vicuna-7b_vicuna_v1_frames_32_stride_2