add tiny bert
Browse files- config.json +19 -0
- model.safetensors +3 -0
- tokenizer_config.json +10 -0
- vocab.txt +100 -0
config.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"model_type": "bert",
|
| 6 |
+
"hidden_size": 32,
|
| 7 |
+
"num_hidden_layers": 2,
|
| 8 |
+
"num_attention_heads": 2,
|
| 9 |
+
"intermediate_size": 64,
|
| 10 |
+
"vocab_size": 100,
|
| 11 |
+
"max_position_embeddings": 64,
|
| 12 |
+
"type_vocab_size": 2,
|
| 13 |
+
"pad_token_id": 0,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"hidden_act": "gelu",
|
| 16 |
+
"attention_probs_dropout_prob": 0.1,
|
| 17 |
+
"hidden_dropout_prob": 0.1,
|
| 18 |
+
"initializer_range": 0.02
|
| 19 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf591418ec1db8794346c5bc5db2713935df4e8de18ddfe270af1cf632cc12ac
|
| 3 |
+
size 116976
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_max_length": 64,
|
| 3 |
+
"do_lower_case": true,
|
| 4 |
+
"tokenizer_class": "BertTokenizer",
|
| 5 |
+
"unk_token": "[UNK]",
|
| 6 |
+
"pad_token": "[PAD]",
|
| 7 |
+
"cls_token": "[CLS]",
|
| 8 |
+
"sep_token": "[SEP]",
|
| 9 |
+
"mask_token": "[MASK]"
|
| 10 |
+
}
|
vocab.txt
ADDED
|
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[PAD]
|
| 2 |
+
[UNK]
|
| 3 |
+
[CLS]
|
| 4 |
+
[SEP]
|
| 5 |
+
[MASK]
|
| 6 |
+
tok0
|
| 7 |
+
tok1
|
| 8 |
+
tok2
|
| 9 |
+
tok3
|
| 10 |
+
tok4
|
| 11 |
+
tok5
|
| 12 |
+
tok6
|
| 13 |
+
tok7
|
| 14 |
+
tok8
|
| 15 |
+
tok9
|
| 16 |
+
tok10
|
| 17 |
+
tok11
|
| 18 |
+
tok12
|
| 19 |
+
tok13
|
| 20 |
+
tok14
|
| 21 |
+
tok15
|
| 22 |
+
tok16
|
| 23 |
+
tok17
|
| 24 |
+
tok18
|
| 25 |
+
tok19
|
| 26 |
+
tok20
|
| 27 |
+
tok21
|
| 28 |
+
tok22
|
| 29 |
+
tok23
|
| 30 |
+
tok24
|
| 31 |
+
tok25
|
| 32 |
+
tok26
|
| 33 |
+
tok27
|
| 34 |
+
tok28
|
| 35 |
+
tok29
|
| 36 |
+
tok30
|
| 37 |
+
tok31
|
| 38 |
+
tok32
|
| 39 |
+
tok33
|
| 40 |
+
tok34
|
| 41 |
+
tok35
|
| 42 |
+
tok36
|
| 43 |
+
tok37
|
| 44 |
+
tok38
|
| 45 |
+
tok39
|
| 46 |
+
tok40
|
| 47 |
+
tok41
|
| 48 |
+
tok42
|
| 49 |
+
tok43
|
| 50 |
+
tok44
|
| 51 |
+
tok45
|
| 52 |
+
tok46
|
| 53 |
+
tok47
|
| 54 |
+
tok48
|
| 55 |
+
tok49
|
| 56 |
+
tok50
|
| 57 |
+
tok51
|
| 58 |
+
tok52
|
| 59 |
+
tok53
|
| 60 |
+
tok54
|
| 61 |
+
tok55
|
| 62 |
+
tok56
|
| 63 |
+
tok57
|
| 64 |
+
tok58
|
| 65 |
+
tok59
|
| 66 |
+
tok60
|
| 67 |
+
tok61
|
| 68 |
+
tok62
|
| 69 |
+
tok63
|
| 70 |
+
tok64
|
| 71 |
+
tok65
|
| 72 |
+
tok66
|
| 73 |
+
tok67
|
| 74 |
+
tok68
|
| 75 |
+
tok69
|
| 76 |
+
tok70
|
| 77 |
+
tok71
|
| 78 |
+
tok72
|
| 79 |
+
tok73
|
| 80 |
+
tok74
|
| 81 |
+
tok75
|
| 82 |
+
tok76
|
| 83 |
+
tok77
|
| 84 |
+
tok78
|
| 85 |
+
tok79
|
| 86 |
+
tok80
|
| 87 |
+
tok81
|
| 88 |
+
tok82
|
| 89 |
+
tok83
|
| 90 |
+
tok84
|
| 91 |
+
tok85
|
| 92 |
+
tok86
|
| 93 |
+
tok87
|
| 94 |
+
tok88
|
| 95 |
+
tok89
|
| 96 |
+
tok90
|
| 97 |
+
tok91
|
| 98 |
+
tok92
|
| 99 |
+
tok93
|
| 100 |
+
tok94
|