davanstrien's picture
davanstrien HF staff
Training in progress, step 500
9978de1
raw
history blame
2.4 kB
{
"_name_or_path": "microsoft/deberta-v3-base",
"architectures": [
"DebertaV2ForTokenClassification"
],
"attention_probs_dropout_prob": 0.1,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 768,
"id2label": {
"0": "L-COLOR",
"1": "U-FOOD",
"2": "U-QUANTITY",
"3": "U-UNIT",
"4": "U-PHYSICAL_QUALITY",
"5": "B-PHYSICAL_QUALITY",
"6": "B-PART",
"7": "I-QUANTITY",
"8": "B-UNIT",
"9": "L-PART",
"10": "B-TASTE",
"11": "B-PURPOSE",
"12": "I-PART",
"13": "L-UNIT",
"14": "L-TASTE",
"15": "L-PHYSICAL_QUALITY",
"16": "L-PROCESS",
"17": "B-FOOD",
"18": "L-QUANTITY",
"19": "I-PHYSICAL_QUALITY",
"20": "B-COLOR",
"21": "I-TASTE",
"22": "B-QUANTITY",
"23": "L-FOOD",
"24": "U-PROCESS",
"25": "L-PURPOSE",
"26": "B-PROCESS",
"27": "I-FOOD",
"28": "U-COLOR",
"29": "U-PART",
"30": "I-PURPOSE",
"31": "I-UNIT",
"32": "I-PROCESS",
"33": "O",
"34": "U-TASTE"
},
"initializer_range": 0.02,
"intermediate_size": 3072,
"label2id": {
"B-COLOR": 20,
"B-FOOD": 17,
"B-PART": 6,
"B-PHYSICAL_QUALITY": 5,
"B-PROCESS": 26,
"B-PURPOSE": 11,
"B-QUANTITY": 22,
"B-TASTE": 10,
"B-UNIT": 8,
"I-FOOD": 27,
"I-PART": 12,
"I-PHYSICAL_QUALITY": 19,
"I-PROCESS": 32,
"I-PURPOSE": 30,
"I-QUANTITY": 7,
"I-TASTE": 21,
"I-UNIT": 31,
"L-COLOR": 0,
"L-FOOD": 23,
"L-PART": 9,
"L-PHYSICAL_QUALITY": 15,
"L-PROCESS": 16,
"L-PURPOSE": 25,
"L-QUANTITY": 18,
"L-TASTE": 14,
"L-UNIT": 13,
"O": 33,
"U-COLOR": 28,
"U-FOOD": 1,
"U-PART": 29,
"U-PHYSICAL_QUALITY": 4,
"U-PROCESS": 24,
"U-QUANTITY": 2,
"U-TASTE": 34,
"U-UNIT": 3
},
"layer_norm_eps": 1e-07,
"max_position_embeddings": 512,
"max_relative_positions": -1,
"model_type": "deberta-v2",
"norm_rel_ebd": "layer_norm",
"num_attention_heads": 12,
"num_hidden_layers": 12,
"pad_token_id": 0,
"pooler_dropout": 0,
"pooler_hidden_act": "gelu",
"pooler_hidden_size": 768,
"pos_att_type": [
"p2c",
"c2p"
],
"position_biased_input": false,
"position_buckets": 256,
"relative_attention": true,
"share_att_key": true,
"torch_dtype": "float32",
"transformers_version": "4.21.0",
"type_vocab_size": 0,
"vocab_size": 128100
}