| { | |
| "activation": "gelu", | |
| "architectures": [ | |
| "DistilBertForSequenceClassification" | |
| ], | |
| "attention_dropout": 0.1, | |
| "bos_token_id": null, | |
| "dim": 768, | |
| "dropout": 0.1, | |
| "dtype": "float32", | |
| "eos_token_id": null, | |
| "hidden_dim": 3072, | |
| "id2label": { | |
| "0": "question", | |
| "1": "code", | |
| "2": "creative", | |
| "3": "chat", | |
| "4": "math" | |
| }, | |
| "initializer_range": 0.02, | |
| "label2id": { | |
| "chat": 3, | |
| "code": 1, | |
| "creative": 2, | |
| "math": 4, | |
| "question": 0 | |
| }, | |
| "max_position_embeddings": 512, | |
| "model_type": "distilbert", | |
| "n_heads": 12, | |
| "n_layers": 6, | |
| "pad_token_id": 0, | |
| "problem_type": "single_label_classification", | |
| "qa_dropout": 0.1, | |
| "seq_classif_dropout": 0.2, | |
| "sinusoidal_pos_embds": false, | |
| "tie_weights_": true, | |
| "tie_word_embeddings": true, | |
| "transformers_version": "5.0.0", | |
| "use_cache": false, | |
| "vocab_size": 30522 | |
| } | |