Upload BigBirdForQuestionAnswering
Browse files- config.json +34 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Models/qa_model",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BigBirdForQuestionAnswering"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"attention_type": "block_sparse",
|
| 8 |
+
"block_size": 64,
|
| 9 |
+
"bos_token_id": 1,
|
| 10 |
+
"classifier_dropout": null,
|
| 11 |
+
"eos_token_id": 2,
|
| 12 |
+
"gradient_checkpointing": false,
|
| 13 |
+
"hidden_act": "gelu_new",
|
| 14 |
+
"hidden_dropout_prob": 0.1,
|
| 15 |
+
"hidden_size": 768,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"layer_norm_eps": 1e-12,
|
| 19 |
+
"max_position_embeddings": 4096,
|
| 20 |
+
"model_type": "big_bird",
|
| 21 |
+
"num_attention_heads": 12,
|
| 22 |
+
"num_hidden_layers": 12,
|
| 23 |
+
"num_random_blocks": 3,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"position_embedding_type": "absolute",
|
| 26 |
+
"rescale_embeddings": false,
|
| 27 |
+
"sep_token_id": 66,
|
| 28 |
+
"torch_dtype": "float32",
|
| 29 |
+
"transformers_version": "4.35.1",
|
| 30 |
+
"type_vocab_size": 16,
|
| 31 |
+
"use_bias": true,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 50358
|
| 34 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:941799157512d402dea70c4cf487c843aca19f8ea33d091f18126e60abda2e5a
|
| 3 |
+
size 526481896
|