| { | |
| "_name_or_path": "../artifacts/best_models/rte/L-6_H-256_A-8/student-ckpt", | |
| "architectures": [ | |
| "BertForSequenceClassification" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "classifier_dropout": null, | |
| "finetuning_task": "rte", | |
| "gradient_checkpointing": false, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "hidden_size": 256, | |
| "id2label": { | |
| "0": 0, | |
| "1": 1 | |
| }, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 1024, | |
| "label2id": { | |
| "0": 0, | |
| "1": 1 | |
| }, | |
| "layer_norm_eps": 1e-12, | |
| "max_position_embeddings": 512, | |
| "model_type": "bert", | |
| "num_attention_heads": 8, | |
| "num_hidden_layers": 6, | |
| "pad_token_id": 0, | |
| "position_embedding_type": "absolute", | |
| "problem_type": "single_label_classification", | |
| "transformers_version": "4.18.0", | |
| "type_vocab_size": 2, | |
| "use_cache": true, | |
| "vocab_size": 30522 | |
| } | |