vazish's picture
hyperparam optimization
5add30f unverified
raw
history blame
1.37 kB
{
"_attn_implementation_autoset": true,
"_name_or_path": "/content/cross-encoder/ms-marco-TinyBERT-L-2-v2",
"architectures": [
"BertForSequenceClassification"
],
"attention_probs_dropout_prob": 0.1,
"classifier_dropout": null,
"gradient_checkpointing": false,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 128,
"id2label": {
"0": "cc-csc",
"1": "cc-exp",
"2": "cc-exp-month",
"3": "cc-exp-year",
"4": "cc-name",
"5": "cc-number",
"6": "cc-type",
"7": "email",
"8": "given-name",
"9": "last-name",
"10": "other",
"11": "postal-code"
},
"initializer_range": 0.02,
"intermediate_size": 512,
"label2id": {
"cc-csc": 0,
"cc-exp": 1,
"cc-exp-month": 2,
"cc-exp-year": 3,
"cc-name": 4,
"cc-number": 5,
"cc-type": 6,
"email": 7,
"given-name": 8,
"last-name": 9,
"other": 10,
"postal-code": 11
},
"layer_norm_eps": 1e-12,
"max_position_embeddings": 512,
"model_type": "bert",
"num_attention_heads": 2,
"num_hidden_layers": 2,
"pad_token_id": 0,
"position_embedding_type": "absolute",
"problem_type": "single_label_classification",
"sbert_ce_default_activation_function": "torch.nn.modules.linear.Identity",
"transformers_version": "4.46.3",
"type_vocab_size": 2,
"use_cache": true,
"vocab_size": 30522
}