Instructions to use pmarquees/succinct-router with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use pmarquees/succinct-router with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir succinct-router pmarquees/succinct-router
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
| { | |
| "schema_version": "1", | |
| "architecture": "succinct-router-decoder-v1", | |
| "model": { | |
| "vocab_size": 8192, | |
| "max_sequence_length": 1024, | |
| "layers": 6, | |
| "model_width": 384, | |
| "attention_heads": 6, | |
| "ffn_width": 1024, | |
| "dropout": 0.1, | |
| "candidate_count": 3 | |
| }, | |
| "route_state": "final_<ROUTE>_token", | |
| "tied_lm_head": true | |
| } | |