mehdidn's picture
Training in progress, step 500
4822632
raw
history blame
1.37 kB
{
"_name_or_path": "HooshvareLab/distilbert-fa-zwnj-base",
"activation": "gelu",
"architectures": [
"DistilBertForTokenClassification"
],
"attention_dropout": 0.1,
"dim": 768,
"dropout": 0.1,
"hidden_dim": 3072,
"id2label": {
"0": "O",
"1": "B-DAT",
"2": "I-DAT",
"3": "B-EVE",
"4": "I-EVE",
"5": "B-FAC",
"6": "I-FAC",
"7": "B-LOC",
"8": "I-LOC",
"9": "B-MON",
"10": "I-MON",
"11": "B-ORG",
"12": "I-ORG",
"13": "B-PCT",
"14": "I-PCT",
"15": "B-PER",
"16": "I-PER",
"17": "B-PRO",
"18": "I-PRO",
"19": "B-TIM",
"20": "I-TIM"
},
"initializer_range": 0.02,
"label2id": {
"B-DAT": 1,
"B-EVE": 3,
"B-FAC": 5,
"B-LOC": 7,
"B-MON": 9,
"B-ORG": 11,
"B-PCT": 13,
"B-PER": 15,
"B-PRO": 17,
"B-TIM": 19,
"I-DAT": 2,
"I-EVE": 4,
"I-FAC": 6,
"I-LOC": 8,
"I-MON": 10,
"I-ORG": 12,
"I-PCT": 14,
"I-PER": 16,
"I-PRO": 18,
"I-TIM": 20,
"O": 0
},
"max_position_embeddings": 512,
"model_type": "distilbert",
"n_heads": 12,
"n_layers": 6,
"output_past": true,
"pad_token_id": 0,
"qa_dropout": 0.1,
"seq_classif_dropout": 0.2,
"sinusoidal_pos_embds": false,
"tie_weights_": true,
"torch_dtype": "float32",
"transformers_version": "4.21.2",
"vocab_size": 42000
}