Prettify config
Browse files- config.json +39 -1
config.json
CHANGED
|
@@ -1 +1,39 @@
|
|
| 1 |
-
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GptBertForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"auto_map": {
|
| 6 |
+
"AutoConfig": "configuration_gptbert.GptBertConfig",
|
| 7 |
+
"AutoModel": "modeling_gptbert.GptBertModel",
|
| 8 |
+
"AutoModelForCausalLM": "modeling_gptbert.GptBertForCausalLM",
|
| 9 |
+
"AutoModelForMaskedLM": "modeling_gptbert.GptBertForMaskedLM",
|
| 10 |
+
"AutoModelForSequenceClassification": "modeling_gptbert.GptBertForSequenceClassification",
|
| 11 |
+
"AutoModelForTokenClassification": "modeling_gptbert.GptBertForTokenClassification",
|
| 12 |
+
"AutoModelForQuestionAnswering": "modeling_gptbert.GptBertForQuestionAnswering",
|
| 13 |
+
"AutoModelForMultipleChoice": "modeling_gptbert.GptBertForMultipleChoice"
|
| 14 |
+
},
|
| 15 |
+
"unk_token_id": 1,
|
| 16 |
+
"bos_token_id": 2,
|
| 17 |
+
"eos_token_id": 3,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"mask_token_id": 4,
|
| 20 |
+
"hidden_size": 640,
|
| 21 |
+
"intermediate_size": 1664,
|
| 22 |
+
"max_sequence_length": 16384,
|
| 23 |
+
"num_layers": 24,
|
| 24 |
+
"attention_dropout": 0.0,
|
| 25 |
+
"hidden_dropout": 0.0,
|
| 26 |
+
"embedding_dropout": 0.1,
|
| 27 |
+
"classifier_dropout": 0.2,
|
| 28 |
+
"layer_norm_eps": 1e-07,
|
| 29 |
+
"query_key_head_size": 64,
|
| 30 |
+
"value_head_size": 64,
|
| 31 |
+
"num_attention_heads": 10,
|
| 32 |
+
"rope_theta": 160000,
|
| 33 |
+
"vocab_size": 32768,
|
| 34 |
+
"local_global_ratio": 4,
|
| 35 |
+
"global_window_length": 8192,
|
| 36 |
+
"local_window_length": 256,
|
| 37 |
+
"deterministic_flash_attn": false,
|
| 38 |
+
"use_cache": false
|
| 39 |
+
}
|