Upload config.json with huggingface_hub
Browse files- config.json +7 -7
config.json
CHANGED
|
@@ -4,16 +4,17 @@
|
|
| 4 |
],
|
| 5 |
"attention_probs_dropout_prob": 0.1,
|
| 6 |
"classifier_dropout": null,
|
|
|
|
| 7 |
"export_model_type": "transformer",
|
| 8 |
"gradient_checkpointing": false,
|
| 9 |
"hidden_act": "gelu",
|
| 10 |
"hidden_dropout_prob": 0.1,
|
| 11 |
-
"hidden_size":
|
| 12 |
"id2label": {
|
| 13 |
"0": "LABEL_0"
|
| 14 |
},
|
| 15 |
"initializer_range": 0.02,
|
| 16 |
-
"intermediate_size":
|
| 17 |
"label2id": {
|
| 18 |
"LABEL_0": 0
|
| 19 |
},
|
|
@@ -24,7 +25,7 @@
|
|
| 24 |
"auto_cast": "matmul",
|
| 25 |
"auto_cast_type": "bf16",
|
| 26 |
"compiler_type": "neuronx-cc",
|
| 27 |
-
"compiler_version": "2.21.
|
| 28 |
"disable_fallback": false,
|
| 29 |
"disable_fast_relayout": false,
|
| 30 |
"dynamic_batch_size": false,
|
|
@@ -41,13 +42,12 @@
|
|
| 41 |
"task": "feature-extraction",
|
| 42 |
"tensor_parallel_size": 1
|
| 43 |
},
|
| 44 |
-
"num_attention_heads":
|
| 45 |
-
"num_hidden_layers":
|
| 46 |
"pad_token_id": 0,
|
| 47 |
"position_embedding_type": "absolute",
|
| 48 |
-
"torch_dtype": "float32",
|
| 49 |
"torchscript": true,
|
| 50 |
-
"transformers_version": "4.
|
| 51 |
"type_vocab_size": 2,
|
| 52 |
"use_cache": true,
|
| 53 |
"vocab_size": 30522
|
|
|
|
| 4 |
],
|
| 5 |
"attention_probs_dropout_prob": 0.1,
|
| 6 |
"classifier_dropout": null,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
"export_model_type": "transformer",
|
| 9 |
"gradient_checkpointing": false,
|
| 10 |
"hidden_act": "gelu",
|
| 11 |
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
"id2label": {
|
| 14 |
"0": "LABEL_0"
|
| 15 |
},
|
| 16 |
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
"label2id": {
|
| 19 |
"LABEL_0": 0
|
| 20 |
},
|
|
|
|
| 25 |
"auto_cast": "matmul",
|
| 26 |
"auto_cast_type": "bf16",
|
| 27 |
"compiler_type": "neuronx-cc",
|
| 28 |
+
"compiler_version": "2.21.33363.0+82129205",
|
| 29 |
"disable_fallback": false,
|
| 30 |
"disable_fast_relayout": false,
|
| 31 |
"dynamic_batch_size": false,
|
|
|
|
| 42 |
"task": "feature-extraction",
|
| 43 |
"tensor_parallel_size": 1
|
| 44 |
},
|
| 45 |
+
"num_attention_heads": 12,
|
| 46 |
+
"num_hidden_layers": 12,
|
| 47 |
"pad_token_id": 0,
|
| 48 |
"position_embedding_type": "absolute",
|
|
|
|
| 49 |
"torchscript": true,
|
| 50 |
+
"transformers_version": "4.57.1",
|
| 51 |
"type_vocab_size": 2,
|
| 52 |
"use_cache": true,
|
| 53 |
"vocab_size": 30522
|