drajend9's picture
Training in progress, epoch 1
ef2bfa5 verified
raw
history blame
1.85 kB
{
"_name_or_path": "allenai/longformer-base-4096",
"architectures": [
"LongformerForTokenClassification"
],
"attention_mode": "longformer",
"attention_probs_dropout_prob": 0.1,
"attention_window": [
512,
512,
512,
512,
512,
512,
512,
512,
512,
512,
512,
512
],
"bos_token_id": 0,
"eos_token_id": 2,
"gradient_checkpointing": false,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 768,
"id2label": {
"0": "O",
"1": "B-BUY_G",
"2": "B-BUY_N",
"3": "B-GSTL",
"4": "B-GT_AMT",
"5": "B-GT_AMTL",
"6": "B-INV_DL",
"7": "B-INV_DT",
"8": "B-INV_L",
"9": "B-INV_NO",
"10": "B-SUPP_G",
"11": "B-SUPP_N",
"12": "I-BUY_G",
"13": "I-BUY_N",
"14": "I-GSTL",
"15": "I-GT_AMT",
"16": "I-GT_AMTL",
"17": "I-INV_DL",
"18": "I-INV_DT",
"19": "I-INV_L",
"20": "I-INV_NO",
"21": "I-SUPP_G",
"22": "I-SUPP_N"
},
"ignore_attention_mask": false,
"initializer_range": 0.02,
"intermediate_size": 3072,
"label2id": {
"B-BUY_G": 1,
"B-BUY_N": 2,
"B-GSTL": 3,
"B-GT_AMT": 4,
"B-GT_AMTL": 5,
"B-INV_DL": 6,
"B-INV_DT": 7,
"B-INV_L": 8,
"B-INV_NO": 9,
"B-SUPP_G": 10,
"B-SUPP_N": 11,
"I-BUY_G": 12,
"I-BUY_N": 13,
"I-GSTL": 14,
"I-GT_AMT": 15,
"I-GT_AMTL": 16,
"I-INV_DL": 17,
"I-INV_DT": 18,
"I-INV_L": 19,
"I-INV_NO": 20,
"I-SUPP_G": 21,
"I-SUPP_N": 22,
"O": 0
},
"layer_norm_eps": 1e-05,
"max_position_embeddings": 4098,
"model_type": "longformer",
"num_attention_heads": 12,
"num_hidden_layers": 12,
"onnx_export": false,
"pad_token_id": 1,
"sep_token_id": 2,
"torch_dtype": "float32",
"transformers_version": "4.42.3",
"type_vocab_size": 1,
"vocab_size": 50265
}