MLX
Joblib
Safetensors
English
reasoning
chain-of-thought
context-compression
soft-prompt
apple-silicon
Instructions to use baya1116/hypernet-sp-distill with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use baya1116/hypernet-sp-distill with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir hypernet-sp-distill baya1116/hypernet-sp-distill
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
File size: 1,099 Bytes
b5989f0 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 | {
"data": {
"gate_pts_held": 22589,
"pos_held": 3227,
"idx_held": 3227
},
"A_gate_operating_point": {
"auc": 0.9537,
"best_f1": {
"f1": 0.725,
"precision": 0.674,
"recall": 0.785,
"thresh": 0.696
},
"recall_at_p0.9": 0.359,
"precision_at_r0.9": 0.526,
"shipped_threshold": {
"thresh": 4.04,
"precision": 0.866,
"recall": 0.469
}
},
"B_gate_model_search": {
"logistic_auc": 0.95,
"mlp256_auc": 0.9607,
"mlp512_auc": 0.9604
},
"C_backward_scan": {
"tau=-32.12": {
"reach_rate": 0.755,
"avg_blocks_scanned": 7.18
},
"tau=-31.70": {
"reach_rate": 0.616,
"avg_blocks_scanned": 7.22
},
"tau=-31.32": {
"reach_rate": 0.445,
"avg_blocks_scanned": 7.22
},
"gold_recency_rank": {
"mean": 6.49,
"median": 7,
"frac_in_recent_3": 0.001,
"frac_in_recent_8": 0.747
}
},
"D_bge_small": {
"note": "skipped (block texts not in IX rows); base bge_head is canonical"
},
"final": true,
"elapsed_s": 340.4
} |