updated
Browse files- config.json +4 -10
config.json
CHANGED
|
@@ -5,12 +5,10 @@
|
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
| 7 |
"bos_token_id": 0,
|
| 8 |
-
"classifier_dropout": null,
|
| 9 |
"eos_token_id": 2,
|
| 10 |
-
"gradient_checkpointing": false,
|
| 11 |
"hidden_act": "gelu",
|
| 12 |
"hidden_dropout_prob": 0.1,
|
| 13 |
-
"hidden_size":
|
| 14 |
"id2label": {
|
| 15 |
"0": "O",
|
| 16 |
"1": "B-PER",
|
|
@@ -23,7 +21,7 @@
|
|
| 23 |
"8": "I-MISC"
|
| 24 |
},
|
| 25 |
"initializer_range": 0.02,
|
| 26 |
-
"intermediate_size":
|
| 27 |
"label2id": {
|
| 28 |
"O": 0,
|
| 29 |
"B-PER": 1,
|
|
@@ -38,14 +36,10 @@
|
|
| 38 |
"layer_norm_eps": 1e-05,
|
| 39 |
"max_position_embeddings": 514,
|
| 40 |
"model_type": "xlm-roberta",
|
| 41 |
-
"num_attention_heads":
|
| 42 |
-
"num_hidden_layers":
|
| 43 |
"output_past": true,
|
| 44 |
"pad_token_id": 1,
|
| 45 |
-
"position_embedding_type": "absolute",
|
| 46 |
-
"torch_dtype": "float32",
|
| 47 |
-
"transformers_version": "4.13.0",
|
| 48 |
"type_vocab_size": 1,
|
| 49 |
-
"use_cache": true,
|
| 50 |
"vocab_size": 250002
|
| 51 |
}
|
|
|
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
| 7 |
"bos_token_id": 0,
|
|
|
|
| 8 |
"eos_token_id": 2,
|
|
|
|
| 9 |
"hidden_act": "gelu",
|
| 10 |
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 1024,
|
| 12 |
"id2label": {
|
| 13 |
"0": "O",
|
| 14 |
"1": "B-PER",
|
|
|
|
| 21 |
"8": "I-MISC"
|
| 22 |
},
|
| 23 |
"initializer_range": 0.02,
|
| 24 |
+
"intermediate_size": 4096,
|
| 25 |
"label2id": {
|
| 26 |
"O": 0,
|
| 27 |
"B-PER": 1,
|
|
|
|
| 36 |
"layer_norm_eps": 1e-05,
|
| 37 |
"max_position_embeddings": 514,
|
| 38 |
"model_type": "xlm-roberta",
|
| 39 |
+
"num_attention_heads": 16,
|
| 40 |
+
"num_hidden_layers": 24,
|
| 41 |
"output_past": true,
|
| 42 |
"pad_token_id": 1,
|
|
|
|
|
|
|
|
|
|
| 43 |
"type_vocab_size": 1,
|
|
|
|
| 44 |
"vocab_size": 250002
|
| 45 |
}
|