|
|
--- |
|
|
base_model: tencent/HunyuanVideo-Foley |
|
|
tags: |
|
|
- quantized |
|
|
- fp8 |
|
|
- audio-generation |
|
|
- video-to-audio |
|
|
- comfyui |
|
|
library_name: transformers |
|
|
--- |
|
|
|
|
|
# HunyuanVideo-Foley FP8 Quantized |
|
|
|
|
|
This is an FP8 quantized version of [tencent/HunyuanVideo-Foley](https://huggingface.co/tencent/HunyuanVideo-Foley) optimized for reduced VRAM usage while maintaining audio generation quality. |
|
|
|
|
|
## Quantization Details |
|
|
|
|
|
- **Quantization Method**: FP8 E5M2 & E4M3FN weight-only quantization |
|
|
- **Layers Quantized**: Transformer block weights only (attention and FFN layers) |
|
|
- **Preserved Precision**: Normalization layers, embeddings, and biases remain in original precision |
|
|
- **Expected VRAM Savings**: ~30-40% reduction compared to BF16 original |
|
|
- **Memory Usage**: Enables running on <12GB GPUs when combined with other optimizations |
|
|
|
|
|
## Usage |
|
|
|
|
|
### ComfyUI (Recommended) |
|
|
|
|
|
This model is specifically optimized for use with the [ComfyUI-HunyuanVideo-Foley](https://github.com/phazei/ComfyUI-HunyuanVideo-Foley) custom node, which provides: |
|
|
|
|
|
- **VRAM-friendly loading** with ping-pong memory management |
|
|
- **Built-in FP8 support** that automatically handles the quantized weights |
|
|
- **Torch compile integration** for ~30% speed improvements after first run |
|
|
- **Text-to-Audio and Video-to-Audio** modes |
|
|
- **Batch generation** with audio selection tools |
|
|
|
|
|
**Installation:** |
|
|
1. Install the ComfyUI node: [ComfyUI-HunyuanVideo-Foley](https://github.com/phazei/ComfyUI-HunyuanVideo-Foley) |
|
|
2. Download this quantized model to `ComfyUI/models/foley/` |
|
|
3. Enjoy <8GB VRAM usage with high-quality audio generation |
|
|
|
|
|
**Typical VRAM Usage (5s audio, 50 steps):** |
|
|
- Baseline (BF16): ~10-12 GB |
|
|
- With FP8 quantization: ~8-10 GB |
|
|
- Perfect for RTX 3080/4070 Ti and similar GPUs |
|
|
|
|
|
### Other Frameworks |
|
|
|
|
|
The FP8 weights can be used with any framework that supports automatic upcasting of FP8 to FP16/BF16 during computation. The quantized weights maintain compatibility with the original model architecture. |
|
|
|
|
|
## Files |
|
|
|
|
|
- `hunyuanvideo_foley_fp8_e4m3fn.safetensors` - Main model weights in FP8 format |
|
|
|
|
|
## Performance Notes |
|
|
|
|
|
- **Quality**: Maintains comparable audio generation quality to the original model |
|
|
- **Speed**: Conversion overhead is minimal; actual generation speed depends on compute precision |
|
|
- **Memory**: Significant VRAM reduction makes the model accessible on consumer GPUs |
|
|
- **Compatibility**: Drop-in replacement for the original model weights |
|
|
|
|
|
## Original Model |
|
|
|
|
|
This quantization is based on [tencent/HunyuanVideo-Foley](https://huggingface.co/tencent/HunyuanVideo-Foley). Please refer to the original repository for: |
|
|
- Model architecture details |
|
|
- Training information |
|
|
- License terms |
|
|
- Citation information |
|
|
|
|
|
## Technical Details |
|
|
|
|
|
The quantization uses a conservative approach that only converts transformer block weights while preserving precision-sensitive components: |
|
|
- ✅ **Converted**: Attention and FFN layer weights in transformer blocks |
|
|
- ❌ **Preserved**: Normalization layers, embeddings, projections, bias terms |
|
|
|
|
|
This selective quantization strategy maintains model quality while maximizing memory savings. |
