Hugging Face's logo

ArtemVazhentsev
/
gatortron-base-mimic

Text Classification
Transformers
Safetensors
megatron-bert
Inference Endpoints
Model card Files Community
ArtemVazhentsev commited on
Commit
37762d4
·
verified ·
1 Parent(s): 3a5f49a

Upload MegatronBertForSequenceClassification

Browse files
Files changed (2) hide show
  1. config.json +130 -0
  2. model.safetensors +3 -0
config.json ADDED
@@ -0,0 +1,130 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/mnt/s3-data/models/ue_models/UE_Transformer/checkpoint-60000",
3
+ "architectures": [
4
+ "MegatronBertForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "hidden_act": "gelu",
8
+ "hidden_dropout_prob": 0.1,
9
+ "hidden_size": 1024,
10
+ "id2label": {
11
+ "0": "E785",
12
+ "1": "I10",
13
+ "2": "Z87891",
14
+ "3": "K219",
15
+ "4": "F329",
16
+ "5": "I2510",
17
+ "6": "N179",
18
+ "7": "F419",
19
+ "8": "Z7901",
20
+ "9": "Z794",
21
+ "10": "E039",
22
+ "11": "E119",
23
+ "12": "G4733",
24
+ "13": "D649",
25
+ "14": "E669",
26
+ "15": "I4891",
27
+ "16": "F17210",
28
+ "17": "Y929",
29
+ "18": "Z66",
30
+ "19": "J45909",
31
+ "20": "Z7902",
32
+ "21": "J449",
33
+ "22": "D62",
34
+ "23": "02HV33Z",
35
+ "24": "N390",
36
+ "25": "I129",
37
+ "26": "E1122",
38
+ "27": "E871",
39
+ "28": "I252",
40
+ "29": "N189",
41
+ "30": "E872",
42
+ "31": "Z8673",
43
+ "32": "Z955",
44
+ "33": "Z86718",
45
+ "34": "G8929",
46
+ "35": "I110",
47
+ "36": "K5900",
48
+ "37": "N400",
49
+ "38": "N183",
50
+ "39": "I480",
51
+ "40": "I130",
52
+ "41": "G4700",
53
+ "42": "D696",
54
+ "43": "Z951",
55
+ "44": "M109",
56
+ "45": "Y92239",
57
+ "46": "J9601",
58
+ "47": "J189",
59
+ "48": "Z23",
60
+ "49": "Y92230"
61
+ },
62
+ "initializer_range": 0.02,
63
+ "intermediate_size": 4096,
64
+ "label2id": {
65
+ "02HV33Z": "23",
66
+ "D62": "22",
67
+ "D649": "13",
68
+ "D696": "42",
69
+ "E039": "10",
70
+ "E1122": "26",
71
+ "E119": "11",
72
+ "E669": "14",
73
+ "E785": "0",
74
+ "E871": "27",
75
+ "E872": "30",
76
+ "F17210": "16",
77
+ "F329": "4",
78
+ "F419": "7",
79
+ "G4700": "41",
80
+ "G4733": "12",
81
+ "G8929": "34",
82
+ "I10": "1",
83
+ "I110": "35",
84
+ "I129": "25",
85
+ "I130": "40",
86
+ "I2510": "5",
87
+ "I252": "28",
88
+ "I480": "39",
89
+ "I4891": "15",
90
+ "J189": "47",
91
+ "J449": "21",
92
+ "J45909": "19",
93
+ "J9601": "46",
94
+ "K219": "3",
95
+ "K5900": "36",
96
+ "M109": "44",
97
+ "N179": "6",
98
+ "N183": "38",
99
+ "N189": "29",
100
+ "N390": "24",
101
+ "N400": "37",
102
+ "Y92230": "49",
103
+ "Y92239": "45",
104
+ "Y929": "17",
105
+ "Z23": "48",
106
+ "Z66": "18",
107
+ "Z7901": "8",
108
+ "Z7902": "20",
109
+ "Z794": "9",
110
+ "Z86718": "33",
111
+ "Z8673": "31",
112
+ "Z87891": "2",
113
+ "Z951": "43",
114
+ "Z955": "32"
115
+ },
116
+ "layer_norm_eps": 1e-12,
117
+ "max_position_embeddings": 512,
118
+ "model_type": "megatron-bert",
119
+ "num_attention_heads": 16,
120
+ "num_hidden_layers": 24,
121
+ "pad_token_id": 0,
122
+ "position_embedding_type": "absolute",
123
+ "problem_type": "multi_label_classification",
124
+ "tokenizer_type": "BertWordPieceCase",
125
+ "torch_dtype": "float32",
126
+ "transformers_version": "4.36.1",
127
+ "type_vocab_size": 2,
128
+ "use_cache": true,
129
+ "vocab_size": 50176
130
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bddf456011c92a0db8c64c317be3ffb1afaf7a550848587af16b60e2281301d
3
+ size 1421321064