Instructions to use MusicForge/magnet-small-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use MusicForge/magnet-small-4bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download MusicForge/magnet-small-4bit --local-dir magnet-small-4bit
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Download config.json from MusicForge/magnet-small-4bit: direct link, hf CLI and curl.
- Browser
- Download file 741 Bytes
-
https://huggingface.co/MusicForge/magnet-small-4bit/resolve/main/config.json
- Command line
-
hf download hf://MusicForge/magnet-small-4bit/config.json
-
curl -L -o config.json https://huggingface.co/MusicForge/magnet-small-4bit/resolve/main/config.json
741 Bytes
| { | |
| "n_q": 4, | |
| "card": 2048, | |
| "dim": 1024, | |
| "num_heads": 16, | |
| "num_layers": 24, | |
| "ffn_dim": 4096, | |
| "segment_duration": 30, | |
| "frame_rate": 50, | |
| "subcodes_context": 5, | |
| "span_len": 3, | |
| "sample_rate": 32000, | |
| "t5_name": "t5-base", | |
| "t5_dim": 768, | |
| "encodec_name": "encodec-32khz", | |
| "quantization": { | |
| "mode": "mlx_affine", | |
| "bits": 4, | |
| "group_size": 64, | |
| "targets": [ | |
| "self_attn.q_proj.weight", | |
| "self_attn.k_proj.weight", | |
| "self_attn.v_proj.weight", | |
| "self_attn.out_proj.weight", | |
| "cross_attn.q_proj.weight", | |
| "cross_attn.k_proj.weight", | |
| "cross_attn.v_proj.weight", | |
| "cross_attn.out_proj.weight", | |
| "linear1.weight", | |
| "linear2.weight" | |
| ] | |
| }, | |
| "format": "int4" | |
| } |