apps1/without_distillation

1{2  "add_cross_attention": false,3  "architectures": [4    "BertHashForSequenceClassification"5  ],6  "attention_probs_dropout_prob": 0.1,7  "auto_map": {8    "AutoConfig": "configuration_bert_hash.BertHashConfig",9    "AutoModel": "modeling_bert_hash.BertHashModel",10    "AutoModelForMaskedLM": "modeling_bert_hash.BertHashForMaskedLM",11    "AutoModelForSequenceClassification": "modeling_bert_hash.BertHashForSequenceClassification"12  },13  "bos_token_id": null,14  "classifier_dropout": null,15  "dtype": "float32",16  "eos_token_id": null,17  "hidden_act": "gelu",18  "hidden_dropout_prob": 0.1,19  "hidden_size": 128,20  "id2label": {21    "0": "LABEL_0",22    "1": "LABEL_1",23    "2": "LABEL_2"24  },25  "initializer_range": 0.02,26  "intermediate_size": 512,27  "is_decoder": false,28  "label2id": {29    "LABEL_0": 0,30    "LABEL_1": 1,31    "LABEL_2": 232  },33  "layer_norm_eps": 1e-12,34  "max_position_embeddings": 512,35  "model_type": "bert_hash",36  "num_attention_heads": 2,37  "num_hidden_layers": 2,38  "pad_token_id": 0,39  "position_embedding_type": "absolute",40  "projections": 16,41  "tie_word_embeddings": true,42  "transformers_version": "5.9.0",43  "type_vocab_size": 2,44  "use_cache": false,45  "vocab_size": 3052246}47