README.md

---
license: apache-2.0
---
### HunyuanVideo_MLLM_text_encoder
This model is adapted from the `llava-llama-3-8b-v1_1-transformers` model, retaining only its `model.language_model` and `processor.tokenizer`. This organization is not the provider of the original model and has no legal affiliation with it.

Intended for use as a text encoder for the `HunyuanVideo` open-source model.

#### You can download the model using the following git clone command or the ModelScope SDK:

SDK Download
```bash
# Install ModelScope
pip install modelscope
``` 
```python
# Download model via SDK
from modelscope import snapshot_download
model_dir = snapshot_download('DiffSynth-Studio/HunyuanVideo_MLLM_text_encoder')
```

Git Download
```
# Git model download
git clone https://www.modelscope.cn/DiffSynth-Studio/HunyuanVideo_MLLM_text_encoder.git
```