Commit
•
1492f8b
1
Parent(s):
6558ddf
Upload config.json
Browse files- config.json +157 -0
config.json
ADDED
|
@@ -0,0 +1,157 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"vocab_size": 30522,
|
| 3 |
+
"hidden_size": 768,
|
| 4 |
+
"num_hidden_layers": 12,
|
| 5 |
+
"num_attention_heads": 12,
|
| 6 |
+
"intermediate_size": 3072,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.1,
|
| 9 |
+
"attention_probs_dropout_prob": 0.1,
|
| 10 |
+
"max_position_embeddings": 512,
|
| 11 |
+
"type_vocab_size": 2,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layers": [
|
| 14 |
+
{
|
| 15 |
+
"class_name": "InputLayer",
|
| 16 |
+
"config": {
|
| 17 |
+
"batch_input_shape": [
|
| 18 |
+
null,
|
| 19 |
+
160
|
| 20 |
+
],
|
| 21 |
+
"dtype": "float32",
|
| 22 |
+
"sparse": false,
|
| 23 |
+
"ragged": false,
|
| 24 |
+
"name": "input_word_ids"
|
| 25 |
+
}
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"class_name": "InputLayer",
|
| 29 |
+
"config": {
|
| 30 |
+
"batch_input_shape": [
|
| 31 |
+
null,
|
| 32 |
+
160
|
| 33 |
+
],
|
| 34 |
+
"dtype": "float32",
|
| 35 |
+
"sparse": false,
|
| 36 |
+
"ragged": false,
|
| 37 |
+
"name": "input_mask"
|
| 38 |
+
}
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"class_name": "InputLayer",
|
| 42 |
+
"config": {
|
| 43 |
+
"batch_input_shape": [
|
| 44 |
+
null,
|
| 45 |
+
160
|
| 46 |
+
],
|
| 47 |
+
"dtype": "float32",
|
| 48 |
+
"sparse": false,
|
| 49 |
+
"ragged": false,
|
| 50 |
+
"name": "segment_ids"
|
| 51 |
+
}
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"class_name": "KerasLayer",
|
| 55 |
+
"config": {
|
| 56 |
+
"name": "keras_layer",
|
| 57 |
+
"trainable": true,
|
| 58 |
+
"dtype": "float32",
|
| 59 |
+
"layer_config": {
|
| 60 |
+
"class_name": "TFRobertaModel",
|
| 61 |
+
"config": {
|
| 62 |
+
"vocab_size": 30522,
|
| 63 |
+
"hidden_size": 768,
|
| 64 |
+
"num_hidden_layers": 12,
|
| 65 |
+
"num_attention_heads": 12,
|
| 66 |
+
"intermediate_size": 3072,
|
| 67 |
+
"hidden_act": "gelu",
|
| 68 |
+
"hidden_dropout_prob": 0.1,
|
| 69 |
+
"attention_probs_dropout_prob": 0.1,
|
| 70 |
+
"max_position_embeddings": 512,
|
| 71 |
+
"type_vocab_size": 2,
|
| 72 |
+
"initializer_range": 0.02,
|
| 73 |
+
"architectures": [
|
| 74 |
+
"TFRobertaModel"
|
| 75 |
+
]
|
| 76 |
+
}
|
| 77 |
+
},
|
| 78 |
+
"class_name": "TFRobertaModel",
|
| 79 |
+
"config": {
|
| 80 |
+
"vocab_size": 30522,
|
| 81 |
+
"hidden_size": 768,
|
| 82 |
+
"num_hidden_layers": 12,
|
| 83 |
+
"num_attention_heads": 12,
|
| 84 |
+
"intermediate_size": 3072,
|
| 85 |
+
"hidden_act": "gelu",
|
| 86 |
+
"hidden_dropout_prob": 0.1,
|
| 87 |
+
"attention_probs_dropout_prob": 0.1,
|
| 88 |
+
"max_position_embeddings": 512,
|
| 89 |
+
"type_vocab_size": 2,
|
| 90 |
+
"initializer_range": 0.02,
|
| 91 |
+
"architectures": [
|
| 92 |
+
"TFRobertaModel"
|
| 93 |
+
]
|
| 94 |
+
}
|
| 95 |
+
}
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"class_name": "TensorSliceOp",
|
| 99 |
+
"config": {
|
| 100 |
+
"name": "tf.__operators__.getitem",
|
| 101 |
+
"trainable": true,
|
| 102 |
+
"dtype": "float32",
|
| 103 |
+
"axis": 0
|
| 104 |
+
}
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"class_name": "Dense",
|
| 108 |
+
"config": {
|
| 109 |
+
"name": "dense",
|
| 110 |
+
"trainable": true,
|
| 111 |
+
"dtype": "float32",
|
| 112 |
+
"units": 1,
|
| 113 |
+
"activation": "linear",
|
| 114 |
+
"use_bias": true,
|
| 115 |
+
"kernel_initializer": {
|
| 116 |
+
"class_name": "GlorotUniform",
|
| 117 |
+
"config": {
|
| 118 |
+
"seed": null
|
| 119 |
+
}
|
| 120 |
+
},
|
| 121 |
+
"bias_initializer": {
|
| 122 |
+
"class_name": "Zeros",
|
| 123 |
+
"config": {}
|
| 124 |
+
},
|
| 125 |
+
"kernel_regularizer": null,
|
| 126 |
+
"bias_regularizer": null,
|
| 127 |
+
"activity_regularizer": null,
|
| 128 |
+
"kernel_constraint": null,
|
| 129 |
+
"bias_constraint": null
|
| 130 |
+
}
|
| 131 |
+
}
|
| 132 |
+
],
|
| 133 |
+
"input_layers": [
|
| 134 |
+
[
|
| 135 |
+
"input_word_ids",
|
| 136 |
+
0,
|
| 137 |
+
0
|
| 138 |
+
],
|
| 139 |
+
[
|
| 140 |
+
"input_mask",
|
| 141 |
+
0,
|
| 142 |
+
0
|
| 143 |
+
],
|
| 144 |
+
[
|
| 145 |
+
"segment_ids",
|
| 146 |
+
0,
|
| 147 |
+
0
|
| 148 |
+
]
|
| 149 |
+
],
|
| 150 |
+
"output_layers": [
|
| 151 |
+
[
|
| 152 |
+
"dense",
|
| 153 |
+
0,
|
| 154 |
+
0
|
| 155 |
+
]
|
| 156 |
+
]
|
| 157 |
+
}
|