---
license: mit
tags:
- lora
- music-generation
- diffrhythm2
- lemm
library_name: diffusers
---

# LEMM LoRA: metal-data1_v2_0

This is a LoRA (Low-Rank Adaptation) adapter for DiffRhythm2 music generation, trained using LEMM (Let Everyone Make Music).

## About LEMM

LEMM is an advanced AI music generation system that allows you to:
- Generate high-quality music with built-in vocals
- Train custom LoRA adapters for specific styles
- Fine-tune models on your own datasets

🎵 **Try it**: [LEMM Space](https://huggingface.co/spaces/Gamahea/lemm-test-100)

## Training Configuration

- **Dataset**: N/A
- **Epochs**: N/A
- **Learning Rate**: 0.0003
- **Batch Size**: 4
- **LoRA Rank**: 16

## How to Use

### In LEMM Space
1. Visit [LEMM](https://huggingface.co/spaces/Gamahea/lemm-test-100)
2. Go to "LoRA Management" tab
3. Enter this model ID: `Gamahea/lemm-lora-metal-data1_v2_0`
4. Click "Download from Hub"
5. Use in generation or as base for continued training

### In Your Code
```python
from pathlib import Path
from huggingface_hub import snapshot_download

# Download LoRA
lora_path = snapshot_download(
    repo_id="Gamahea/lemm-lora-metal-data1_v2_0",
    local_dir="./loras/metal-data1_v2_0"
)

# Load and use with DiffRhythm2
# (See LEMM documentation for integration)
```

## Model Files

- `final_model.pt` - Trained LoRA weights
- `config.yaml` - Training configuration
- `README.md` - This file

## Collection

Part of the [LEMM 1.0.0 Pre-Beta Collection](https://huggingface.co/collections/Gamahea/lemm-100-pre-beta)

## License

MIT License - Free to use and modify