# Video Captioning

Human labeling of videos is expensive and time-consuming. We adopt powerful image captioning models to generate captions for videos. Although GPT-4V achieves a better performance, its 20s/sample speed is too slow for us. LLaVA is the second best open-source model in [MMMU](https://mmmu-benchmark.github.io/) and accepts any resolution. We find the quality of 34B model is comparable.

![Caption](https://i0.imgs.ovh/2024/03/16/eXdvC.png)

## LLaVA Captioning

We extract three frames from the video for captioning. With batch inference, we can achieve 10 times speedup. With approximatly 720p resolution and 1 frames, the speed is 2~3 videos/s on 8 GPUs. If we resize the smaller side to 336, the speed can be 8 videos/s. In Open-Sora v1.1, to lower the cost, we use the 7B model.

### Requirement

```bash
# create conda env
conda create -n llava python=3.10 -y
conda activate llava

# install torch
pip install torch torchvision

# clone llava
git clone https://github.com/haotian-liu/LLaVA.git
cd LLaVA
# CAUTION: This line is to remove torch dependency in pyproject.toml, which is:
# "torch==2.1.2", "torchvision==0.16.2",
# It is better manually remove it in your local pyproject.toml
sed -i '16d' pyproject.toml

# install llava
pip install --upgrade pip  # enable PEP 660 support
pip install -e .

# install flash attention
pip install flash-attn --no-build-isolation
# install colossalai and decord
pip install colossalai decord
```

### Usage

Prepare a csv file for processing. The csv file can be generated by `convert_dataset.py` according to its [documentation](/tools/datasets/README.md). Then, run the following command to generate captions for videos/images with Llava:

```bash
# caption with mistral-7B
torchrun --nproc_per_node 8 --standalone -m tools.caption.caption_llava DATA.csv --dp-size 8 --tp-size 1 --model-path liuhaotian/llava-v1.6-mistral-7b --prompt video

# caption with llava-34B
# NOTE: remember to enable flash attention for this model
torchrun --nproc_per_node 8 --standalone -m tools.caption.caption_llava DATA.csv --dp-size 4 --tp-size 2 --model-path liuhaotian/llava-v1.6-34b --prompt image-3ex --flash-attention

# we run this on 8xH800 GPUs
torchrun --nproc_per_node 8 --standalone -m tools.caption.caption_llava DATA.csv --tp-size 2 --dp-size 4 --bs 16

# at least two 80G GPUs are required
torchrun --nproc_per_node 2 --standalone -m tools.caption.caption_llava DATA.csv --tp-size 2 --dp-size 1 --bs 16

# can also caption images
torchrun --nproc_per_node 2 --standalone -m tools.caption.caption_llava DATA.csv --tp-size 2 --dp-size 1 --bs 16 --prompt image-3ex
```

Please note that you should add the `--flash-attention` flag when running with Llama-based Llava models as it provides speedup but do turn it off for mistral-based ones. Reasons can be found in [this issue](https://discuss.huggingface.co/t/flash-attention-has-no-effect-on-inference/73453).

After running the script, with `dp-size=N`, you will get `N` parts of csv files. Run the following command to merge them:

```bash
python -m tools.datasets.datautil DATA_caption_part*.csv --output DATA_caption.csv
```

### Resume

Sometimes the process may be interrupted. We can resume the process by running the following command:

```bash
# merge generated results
python -m tools.datasets.datautil DATA_caption_part*.csv --output DATA_caption.csv

# get the remaining videos
python -m tools.datasets.datautil DATA.csv --difference DATA_caption.csv --output DATA_remaining.csv
```

Then use the output csv file to resume the process.

## GPT-4V Captioning

Run the following command to generate captions for videos with GPT-4V:

```bash
# output: DATA_caption.csv
python -m tools.caption.caption_gpt4 DATA.csv --key $OPENAI_API_KEY
```

The cost is approximately $0.01 per video (3 frames per video).