MLX
English
nima_joint_schema
tool-routing
decision-model
frozen-backbone
joint-schema
custom-code
research
Instructions to use neimand/nima with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use neimand/nima with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download neimand/nima --local-dir nima
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Download training_metadata.json from neimand/nima: direct link, hf CLI and curl.
- Browser
- Download file 8.02 kB
-
https://huggingface.co/neimand/nima/resolve/main/training_metadata.json
- Command line
-
hf download hf://neimand/nima/training_metadata.json
-
curl -L -o training_metadata.json https://huggingface.co/neimand/nima/resolve/main/training_metadata.json
8.02 kB
| { | |
| "best_epoch": 1, | |
| "selection_nll": 0.43172744103717203, | |
| "versions": { | |
| "mlx": "0.32.3", | |
| "mlx-lm": "0.32.0", | |
| "numpy": "2.5.3" | |
| }, | |
| "effective_training_rows": 1706, | |
| "optimizer_updates": 17120, | |
| "seconds": 89.4501815829999, | |
| "memory": { | |
| "mlx_peak_allocated_bytes": 87674924, | |
| "mlx_active_allocated_bytes": 21107772, | |
| "process_max_rss_bytes_macos": 1636679680 | |
| }, | |
| "runtime_checks": { | |
| "variable_candidates": [ | |
| 2, | |
| 5, | |
| 8 | |
| ], | |
| "padding_isolation": true, | |
| "candidate_permutation_equivariance": true, | |
| "bidirectional_candidate_interaction": true, | |
| "evidence_influence": true, | |
| "finite_gradients": true, | |
| "gradient_norms": { | |
| "evidence_attention": 1.0483995552985155, | |
| "joint": 3.5755980015137983, | |
| "project": 19.5132737136767, | |
| "score": 7.060013324843847 | |
| }, | |
| "reload_agreement": true, | |
| "trainable_parameters": 1318657 | |
| }, | |
| "validation": { | |
| "selection": { | |
| "count": 256, | |
| "unique_requests": 128, | |
| "accuracy": 0.8359375, | |
| "correct": 214, | |
| "nll": 0.43172744103717203, | |
| "by_tool": { | |
| "web_search": { | |
| "count": 32, | |
| "correct": 32 | |
| }, | |
| "rag_search": { | |
| "count": 32, | |
| "correct": 24 | |
| }, | |
| "database_search": { | |
| "count": 32, | |
| "correct": 32 | |
| }, | |
| "send_email": { | |
| "count": 32, | |
| "correct": 19 | |
| }, | |
| "weather_search": { | |
| "count": 32, | |
| "correct": 31 | |
| }, | |
| "calendar_search": { | |
| "count": 32, | |
| "correct": 32 | |
| }, | |
| "calculator": { | |
| "count": 32, | |
| "correct": 32 | |
| }, | |
| "no_tool": { | |
| "count": 32, | |
| "correct": 12 | |
| } | |
| }, | |
| "confusion": [ | |
| { | |
| "target": "calculator", | |
| "prediction": "calculator", | |
| "count": 32 | |
| }, | |
| { | |
| "target": "calendar_search", | |
| "prediction": "calendar_search", | |
| "count": 32 | |
| }, | |
| { | |
| "target": "database_search", | |
| "prediction": "database_search", | |
| "count": 32 | |
| }, | |
| { | |
| "target": "no_tool", | |
| "prediction": "database_search", | |
| "count": 6 | |
| }, | |
| { | |
| "target": "no_tool", | |
| "prediction": "no_tool", | |
| "count": 12 | |
| }, | |
| { | |
| "target": "no_tool", | |
| "prediction": "send_email", | |
| "count": 10 | |
| }, | |
| { | |
| "target": "no_tool", | |
| "prediction": "weather_search", | |
| "count": 4 | |
| }, | |
| { | |
| "target": "rag_search", | |
| "prediction": "database_search", | |
| "count": 4 | |
| }, | |
| { | |
| "target": "rag_search", | |
| "prediction": "no_tool", | |
| "count": 4 | |
| }, | |
| { | |
| "target": "rag_search", | |
| "prediction": "rag_search", | |
| "count": 24 | |
| }, | |
| { | |
| "target": "send_email", | |
| "prediction": "database_search", | |
| "count": 1 | |
| }, | |
| { | |
| "target": "send_email", | |
| "prediction": "no_tool", | |
| "count": 12 | |
| }, | |
| { | |
| "target": "send_email", | |
| "prediction": "send_email", | |
| "count": 19 | |
| }, | |
| { | |
| "target": "weather_search", | |
| "prediction": "no_tool", | |
| "count": 1 | |
| }, | |
| { | |
| "target": "weather_search", | |
| "prediction": "weather_search", | |
| "count": 31 | |
| }, | |
| { | |
| "target": "web_search", | |
| "prediction": "web_search", | |
| "count": 32 | |
| } | |
| ], | |
| "no_tool": { | |
| "precision": 0.41379310344827586, | |
| "recall": 0.375, | |
| "false_abstentions": 17 | |
| }, | |
| "orders_disagree_requests": 21, | |
| "all_views_correct_requests": 98 | |
| }, | |
| "calibration": { | |
| "count": 128, | |
| "unique_requests": 128, | |
| "accuracy": 0.796875, | |
| "correct": 102, | |
| "nll": 0.5966955877251904, | |
| "by_tool": { | |
| "web_search": { | |
| "count": 16, | |
| "correct": 15 | |
| }, | |
| "rag_search": { | |
| "count": 16, | |
| "correct": 5 | |
| }, | |
| "database_search": { | |
| "count": 16, | |
| "correct": 14 | |
| }, | |
| "send_email": { | |
| "count": 16, | |
| "correct": 9 | |
| }, | |
| "weather_search": { | |
| "count": 16, | |
| "correct": 15 | |
| }, | |
| "calendar_search": { | |
| "count": 16, | |
| "correct": 16 | |
| }, | |
| "calculator": { | |
| "count": 16, | |
| "correct": 16 | |
| }, | |
| "no_tool": { | |
| "count": 16, | |
| "correct": 12 | |
| } | |
| }, | |
| "confusion": [ | |
| { | |
| "target": "calculator", | |
| "prediction": "calculator", | |
| "count": 16 | |
| }, | |
| { | |
| "target": "calendar_search", | |
| "prediction": "calendar_search", | |
| "count": 16 | |
| }, | |
| { | |
| "target": "database_search", | |
| "prediction": "database_search", | |
| "count": 14 | |
| }, | |
| { | |
| "target": "database_search", | |
| "prediction": "no_tool", | |
| "count": 2 | |
| }, | |
| { | |
| "target": "no_tool", | |
| "prediction": "database_search", | |
| "count": 1 | |
| }, | |
| { | |
| "target": "no_tool", | |
| "prediction": "no_tool", | |
| "count": 12 | |
| }, | |
| { | |
| "target": "no_tool", | |
| "prediction": "send_email", | |
| "count": 3 | |
| }, | |
| { | |
| "target": "rag_search", | |
| "prediction": "database_search", | |
| "count": 7 | |
| }, | |
| { | |
| "target": "rag_search", | |
| "prediction": "no_tool", | |
| "count": 4 | |
| }, | |
| { | |
| "target": "rag_search", | |
| "prediction": "rag_search", | |
| "count": 5 | |
| }, | |
| { | |
| "target": "send_email", | |
| "prediction": "no_tool", | |
| "count": 7 | |
| }, | |
| { | |
| "target": "send_email", | |
| "prediction": "send_email", | |
| "count": 9 | |
| }, | |
| { | |
| "target": "weather_search", | |
| "prediction": "no_tool", | |
| "count": 1 | |
| }, | |
| { | |
| "target": "weather_search", | |
| "prediction": "weather_search", | |
| "count": 15 | |
| }, | |
| { | |
| "target": "web_search", | |
| "prediction": "no_tool", | |
| "count": 1 | |
| }, | |
| { | |
| "target": "web_search", | |
| "prediction": "web_search", | |
| "count": 15 | |
| } | |
| ], | |
| "no_tool": { | |
| "precision": 0.4444444444444444, | |
| "recall": 0.75, | |
| "false_abstentions": 15 | |
| }, | |
| "orders_disagree_requests": 0, | |
| "all_views_correct_requests": 102 | |
| } | |
| }, | |
| "head_sha256": "0e8f91e97b1ac2e309463fbb8b0e10c9d165bf4961bd40613aaa0603830e2988", | |
| "calibration_policy": "temperature fixed at 1; no thresholds or calibration fitting in this milestone", | |
| "source_hashes": { | |
| "run.py": "6e8d151b03c387153bbcf229f91ffa3fe1e05f042f1e271ba0aaa22155801f5b", | |
| "validate_dataset.py": "a811abc97d54b8a1ca5e48a2f12c5591d3300685f7faf7e4c2b61161172a7c22", | |
| "test_offline.py": "1cd7a07e215240161a989e35ccddf45029bdc044d67250a0fdf9cc59a32249fa", | |
| "architecture.json": "46918f976e86089108003daf0c7ece7ef4cdaf677e3ae5b8a0dbb8e19989650b", | |
| "head.py": "5f2ddf1b5b45b0483bd62fc5c86e9cba1f283e854548abdad9bcda6a5e47556c", | |
| "protocol.json": "b7376ce11843210d72a2ffe129bf193af2bdf05fdb9bcf8cc74d8f2b6fe161c6", | |
| "CPU_CHECKS.json": "fe15fc54746af9d40ac37fea8c9ecdaa1e32d77ff7aae7fce0819f43acb0cac2", | |
| "workflow.py": "88b481c52a96a725c800a3f6eb40f70ced04c5ccaadefb1f1187e0d296c4eb36", | |
| "build_data.py": "4a19cc1f4175b43dad77eebdc17078565f05b86b21cdd082277ca507f1f6d717", | |
| "evaluate_pair.py": "5d2a55580220ff43bef4224824c5895a8ebd306f707a6e47b7a5463a9b7cf1d3", | |
| "data.py": "bd1bdb996f2d2de4a2c993bd2cdcb37d0c1d975b604f400c8bd21f3d3df93695" | |
| } | |
| } | |