Update tiny models for BigBirdForMultipleChoice
Browse files- config.json +34 -0
- model.safetensors +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +118 -0
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_cross_attention": false,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BigBirdForMultipleChoice"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"attention_type": "block_sparse",
|
| 8 |
+
"block_size": 8,
|
| 9 |
+
"bos_token_id": 1,
|
| 10 |
+
"classifier_dropout": null,
|
| 11 |
+
"dtype": "float32",
|
| 12 |
+
"eos_token_id": 2,
|
| 13 |
+
"hidden_act": "gelu_new",
|
| 14 |
+
"hidden_dropout_prob": 0.1,
|
| 15 |
+
"hidden_size": 32,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 37,
|
| 18 |
+
"is_decoder": false,
|
| 19 |
+
"layer_norm_eps": 1e-12,
|
| 20 |
+
"max_position_embeddings": 256,
|
| 21 |
+
"model_type": "big_bird",
|
| 22 |
+
"num_attention_heads": 4,
|
| 23 |
+
"num_hidden_layers": 2,
|
| 24 |
+
"num_random_blocks": 3,
|
| 25 |
+
"pad_token_id": 0,
|
| 26 |
+
"rescale_embeddings": false,
|
| 27 |
+
"sep_token_id": 66,
|
| 28 |
+
"tie_word_embeddings": true,
|
| 29 |
+
"transformers_version": "5.16.0.dev0",
|
| 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 sha256:9eb21a1226592792fdbbea08d313bdf23169d20c3b8f68e3daae1f8fa8665f2a
|
| 3 |
+
size 6544020
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "</s>",
|
| 5 |
+
"cls_token": "[CLS]",
|
| 6 |
+
"eos_token": "<s>",
|
| 7 |
+
"extra_special_tokens": [
|
| 8 |
+
"<pad>",
|
| 9 |
+
"</s>",
|
| 10 |
+
"<s>",
|
| 11 |
+
"<sep_0>",
|
| 12 |
+
"<sep_1>",
|
| 13 |
+
"<sep_2>",
|
| 14 |
+
"<sep_3>",
|
| 15 |
+
"<sep_4>",
|
| 16 |
+
"<no_saic_raw_sp>",
|
| 17 |
+
"<length_0>",
|
| 18 |
+
"<length_1>",
|
| 19 |
+
"<length_2>",
|
| 20 |
+
"<length_3>",
|
| 21 |
+
"<length_4>",
|
| 22 |
+
"<length_5>",
|
| 23 |
+
"<length_6>",
|
| 24 |
+
"<length_7>",
|
| 25 |
+
"<length_8>",
|
| 26 |
+
"<length_9>",
|
| 27 |
+
"<length_10>",
|
| 28 |
+
"<length_11>",
|
| 29 |
+
"<length_12>",
|
| 30 |
+
"<length_13>",
|
| 31 |
+
"<length_14>",
|
| 32 |
+
"<length_15>",
|
| 33 |
+
"<length_16>",
|
| 34 |
+
"<length_17>",
|
| 35 |
+
"<length_18>",
|
| 36 |
+
"<length_19>",
|
| 37 |
+
"<length_20>",
|
| 38 |
+
"<length_s>",
|
| 39 |
+
"<length_m>",
|
| 40 |
+
"<length_l>",
|
| 41 |
+
"<::::>",
|
| 42 |
+
"<space>",
|
| 43 |
+
"<unused_34>",
|
| 44 |
+
"<unused_35>",
|
| 45 |
+
"<unused_36>",
|
| 46 |
+
"<no_query_sp>",
|
| 47 |
+
"<no_passage_raw_sp>",
|
| 48 |
+
"<unused_39>",
|
| 49 |
+
"<unused_40>",
|
| 50 |
+
"<unused_41>",
|
| 51 |
+
"<unused_42>",
|
| 52 |
+
"<unused_43>",
|
| 53 |
+
"<unused_44>",
|
| 54 |
+
"<unused_45>",
|
| 55 |
+
"<unused_46>",
|
| 56 |
+
"<unused_47>",
|
| 57 |
+
"<unused_48>",
|
| 58 |
+
"<unused_49>",
|
| 59 |
+
"<paraphrasing_0>",
|
| 60 |
+
"<paraphrasing_1>",
|
| 61 |
+
"<paraphrasing_2>",
|
| 62 |
+
"<paraphrasing_3>",
|
| 63 |
+
"<paraphrasing_4>",
|
| 64 |
+
"<paraphrasing_5>",
|
| 65 |
+
"<paraphrasing_6>",
|
| 66 |
+
"<paraphrasing_7>",
|
| 67 |
+
"<paraphrasing_8>",
|
| 68 |
+
"<empty>",
|
| 69 |
+
"<row>",
|
| 70 |
+
"<cn>",
|
| 71 |
+
"<t>",
|
| 72 |
+
"<ans>",
|
| 73 |
+
"[CLS]",
|
| 74 |
+
"[SEP]",
|
| 75 |
+
"[MASK]",
|
| 76 |
+
"<unused_67>",
|
| 77 |
+
"<unused_68>",
|
| 78 |
+
"<unused_69>",
|
| 79 |
+
"<unused_70>",
|
| 80 |
+
"<unused_71>",
|
| 81 |
+
"<unused_72>",
|
| 82 |
+
"<unused_73>",
|
| 83 |
+
"<unused_74>",
|
| 84 |
+
"<unused_75>",
|
| 85 |
+
"<unused_76>",
|
| 86 |
+
"<unused_77>",
|
| 87 |
+
"<unused_78>",
|
| 88 |
+
"<unused_79>",
|
| 89 |
+
"<unused_80>",
|
| 90 |
+
"<unused_81>",
|
| 91 |
+
"<unused_82>",
|
| 92 |
+
"<unused_83>",
|
| 93 |
+
"<unused_84>",
|
| 94 |
+
"<unused_85>",
|
| 95 |
+
"<unused_86>",
|
| 96 |
+
"<unused_87>",
|
| 97 |
+
"<unused_88>",
|
| 98 |
+
"<unused_89>",
|
| 99 |
+
"<unused_90>",
|
| 100 |
+
"<unused_91>",
|
| 101 |
+
"<unused_92>",
|
| 102 |
+
"<unused_93>",
|
| 103 |
+
"<unused_94>",
|
| 104 |
+
"<unused_95>",
|
| 105 |
+
"<unused_96>",
|
| 106 |
+
"<unused_97>",
|
| 107 |
+
"<unused_98>"
|
| 108 |
+
],
|
| 109 |
+
"is_local": true,
|
| 110 |
+
"local_files_only": false,
|
| 111 |
+
"mask_token": "[MASK]",
|
| 112 |
+
"model_max_length": 256,
|
| 113 |
+
"pad_token": "<pad>",
|
| 114 |
+
"sep_token": "[SEP]",
|
| 115 |
+
"tokenizer_class": "BigBirdTokenizer",
|
| 116 |
+
"unk_id": 100,
|
| 117 |
+
"unk_token": "<unk>"
|
| 118 |
+
}
|