| { |
| "architectures": [ |
| "ModernBertModel" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 50281, |
| "classifier_activation": "silu", |
| "classifier_bias": false, |
| "classifier_dropout": 0.0, |
| "classifier_pooling": "mean", |
| "cls_token_id": 50281, |
| "decoder_bias": true, |
| "deterministic_flash_attn": false, |
| "embedding_dropout": 0.0, |
| "eos_token_id": 50282, |
| "global_attn_every_n_layers": 3, |
| "global_rope_theta": 160000.0, |
| "gradient_checkpointing": false, |
| "hidden_activation": "gelu", |
| "hidden_size": 768, |
| "initializer_cutoff_factor": 2.0, |
| "initializer_range": 0.02, |
| "intermediate_size": 1152, |
| "layer_norm_eps": 1e-05, |
| "local_attention": 128, |
| "local_rope_theta": 10000.0, |
| "max_position_embeddings": 8192, |
| "mlp_bias": false, |
| "mlp_dropout": 0.0, |
| "model_type": "modernbert", |
| "norm_bias": false, |
| "norm_eps": 1e-05, |
| "num_attention_heads": 12, |
| "num_hidden_layers": 22, |
| "pad_token_id": 50283, |
| "position_embedding_type": "absolute", |
| "sep_token_id": 50282, |
| "tie_word_embeddings": true, |
| "torch_dtype": "float32", |
| "transformers_version": "4.48.0", |
| "vocab_size": 50369, |
| "dinah": { |
| "opt_token": "[OPT]", |
| "opt_token_id": 50368, |
| "question_types": { |
| "noul": 0, |
| "choice": 1, |
| "score": 2 |
| }, |
| "max_length": 8192, |
| "pad_multiple": 64, |
| "calibration": { |
| "temperature": [ |
| [ |
| 1.3, |
| 1.0, |
| 1.0, |
| 1.0, |
| 1.0, |
| 1.0, |
| 1.0, |
| 1.0 |
| ], |
| [ |
| 1.0, |
| 1.0, |
| 1.4, |
| 1.35, |
| 1.0, |
| 1.0, |
| 1.0, |
| 1.0 |
| ], |
| [ |
| 1.0, |
| 1.0, |
| 1.3, |
| 1.3, |
| 1.0, |
| 1.0, |
| 1.0, |
| 1.0 |
| ] |
| ], |
| "platt": [ |
| 0.7, |
| -0.1 |
| ], |
| "note": "temperature[question_type][min(max(k,2),9)-2] divides the logits (fit on soft-label NLL); platt maps the confidence logit: sigmoid(a*z+b). Fit on half of the typed-decisions test sets (even indices); the other half is held out." |
| } |
| } |
| } |
