Commit
Β·
b85a1d8
1
Parent(s):
382b14f
Added model
Browse files- config.json +215 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +1 -0
- spiece.model +3 -0
- tokenizer_config.json +1 -0
config.json
ADDED
|
@@ -0,0 +1,215 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/mt5-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MT5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 2048,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 768,
|
| 9 |
+
"decoder_start_token_id": 250030,
|
| 10 |
+
"dropout_rate": 0.1,
|
| 11 |
+
"eos_token_id": 1,
|
| 12 |
+
"feed_forward_proj": "gated-gelu",
|
| 13 |
+
"initializer_factor": 1.0,
|
| 14 |
+
"is_encoder_decoder": true,
|
| 15 |
+
"layer_norm_epsilon": 1e-06,
|
| 16 |
+
"length_penalty": 0.6,
|
| 17 |
+
"max_length": 84,
|
| 18 |
+
"model_type": "mt5",
|
| 19 |
+
"num_beams": 4,
|
| 20 |
+
"num_decoder_layers": 12,
|
| 21 |
+
"num_heads": 12,
|
| 22 |
+
"num_layers": 12,
|
| 23 |
+
"output_past": true,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"relative_attention_num_buckets": 32,
|
| 26 |
+
"task_specific_params": {
|
| 27 |
+
"langid_map": {
|
| 28 |
+
"amharic": [
|
| 29 |
+
35,
|
| 30 |
+
"\u2581<extra_id_64>"
|
| 31 |
+
],
|
| 32 |
+
"arabic": [
|
| 33 |
+
4,
|
| 34 |
+
"\u2581<extra_id_95>"
|
| 35 |
+
],
|
| 36 |
+
"azerbaijani": [
|
| 37 |
+
7,
|
| 38 |
+
"\u2581<extra_id_92>"
|
| 39 |
+
],
|
| 40 |
+
"bengali": [
|
| 41 |
+
42,
|
| 42 |
+
"\u2581<extra_id_57>"
|
| 43 |
+
],
|
| 44 |
+
"burmese": [
|
| 45 |
+
33,
|
| 46 |
+
"\u2581<extra_id_66>"
|
| 47 |
+
],
|
| 48 |
+
"chinese_simplified": [
|
| 49 |
+
40,
|
| 50 |
+
"\u2581<extra_id_59>"
|
| 51 |
+
],
|
| 52 |
+
"chinese_traditional": [
|
| 53 |
+
44,
|
| 54 |
+
"\u2581<extra_id_55>"
|
| 55 |
+
],
|
| 56 |
+
"english": [
|
| 57 |
+
30,
|
| 58 |
+
"\u2581<extra_id_69>"
|
| 59 |
+
],
|
| 60 |
+
"french": [
|
| 61 |
+
10,
|
| 62 |
+
"\u2581<extra_id_89>"
|
| 63 |
+
],
|
| 64 |
+
"gujarati": [
|
| 65 |
+
27,
|
| 66 |
+
"\u2581<extra_id_72>"
|
| 67 |
+
],
|
| 68 |
+
"hausa": [
|
| 69 |
+
43,
|
| 70 |
+
"\u2581<extra_id_56>"
|
| 71 |
+
],
|
| 72 |
+
"hindi": [
|
| 73 |
+
21,
|
| 74 |
+
"\u2581<extra_id_78>"
|
| 75 |
+
],
|
| 76 |
+
"igbo": [
|
| 77 |
+
9,
|
| 78 |
+
"\u2581<extra_id_90>"
|
| 79 |
+
],
|
| 80 |
+
"indonesian": [
|
| 81 |
+
1,
|
| 82 |
+
"\u2581<extra_id_98>"
|
| 83 |
+
],
|
| 84 |
+
"japanese": [
|
| 85 |
+
37,
|
| 86 |
+
"\u2581<extra_id_62>"
|
| 87 |
+
],
|
| 88 |
+
"kirundi": [
|
| 89 |
+
0,
|
| 90 |
+
"\u2581<extra_id_99>"
|
| 91 |
+
],
|
| 92 |
+
"korean": [
|
| 93 |
+
29,
|
| 94 |
+
"\u2581<extra_id_70>"
|
| 95 |
+
],
|
| 96 |
+
"kyrgyz": [
|
| 97 |
+
5,
|
| 98 |
+
"\u2581<extra_id_94>"
|
| 99 |
+
],
|
| 100 |
+
"marathi": [
|
| 101 |
+
13,
|
| 102 |
+
"\u2581<extra_id_86>"
|
| 103 |
+
],
|
| 104 |
+
"nepali": [
|
| 105 |
+
20,
|
| 106 |
+
"\u2581<extra_id_79>"
|
| 107 |
+
],
|
| 108 |
+
"oromo": [
|
| 109 |
+
41,
|
| 110 |
+
"\u2581<extra_id_58>"
|
| 111 |
+
],
|
| 112 |
+
"pashto": [
|
| 113 |
+
34,
|
| 114 |
+
"\u2581<extra_id_65>"
|
| 115 |
+
],
|
| 116 |
+
"persian": [
|
| 117 |
+
23,
|
| 118 |
+
"\u2581<extra_id_76>"
|
| 119 |
+
],
|
| 120 |
+
"pidgin": [
|
| 121 |
+
14,
|
| 122 |
+
"\u2581<extra_id_85>"
|
| 123 |
+
],
|
| 124 |
+
"portuguese": [
|
| 125 |
+
39,
|
| 126 |
+
"\u2581<extra_id_60>"
|
| 127 |
+
],
|
| 128 |
+
"punjabi": [
|
| 129 |
+
17,
|
| 130 |
+
"\u2581<extra_id_82>"
|
| 131 |
+
],
|
| 132 |
+
"russian": [
|
| 133 |
+
36,
|
| 134 |
+
"\u2581<extra_id_63>"
|
| 135 |
+
],
|
| 136 |
+
"scottish_gaelic": [
|
| 137 |
+
24,
|
| 138 |
+
"\u2581<extra_id_75>"
|
| 139 |
+
],
|
| 140 |
+
"serbian_cyrillic": [
|
| 141 |
+
28,
|
| 142 |
+
"\u2581<extra_id_71>"
|
| 143 |
+
],
|
| 144 |
+
"serbian_latin": [
|
| 145 |
+
11,
|
| 146 |
+
"\u2581<extra_id_88>"
|
| 147 |
+
],
|
| 148 |
+
"sinhala": [
|
| 149 |
+
31,
|
| 150 |
+
"\u2581<extra_id_68>"
|
| 151 |
+
],
|
| 152 |
+
"somali": [
|
| 153 |
+
19,
|
| 154 |
+
"\u2581<extra_id_80>"
|
| 155 |
+
],
|
| 156 |
+
"spanish": [
|
| 157 |
+
3,
|
| 158 |
+
"\u2581<extra_id_96>"
|
| 159 |
+
],
|
| 160 |
+
"swahili": [
|
| 161 |
+
18,
|
| 162 |
+
"\u2581<extra_id_81>"
|
| 163 |
+
],
|
| 164 |
+
"tamil": [
|
| 165 |
+
32,
|
| 166 |
+
"\u2581<extra_id_67>"
|
| 167 |
+
],
|
| 168 |
+
"telugu": [
|
| 169 |
+
22,
|
| 170 |
+
"\u2581<extra_id_77>"
|
| 171 |
+
],
|
| 172 |
+
"thai": [
|
| 173 |
+
6,
|
| 174 |
+
"\u2581<extra_id_93>"
|
| 175 |
+
],
|
| 176 |
+
"tigrinya": [
|
| 177 |
+
16,
|
| 178 |
+
"\u2581<extra_id_83>"
|
| 179 |
+
],
|
| 180 |
+
"turkish": [
|
| 181 |
+
15,
|
| 182 |
+
"\u2581<extra_id_84>"
|
| 183 |
+
],
|
| 184 |
+
"ukrainian": [
|
| 185 |
+
2,
|
| 186 |
+
"\u2581<extra_id_97>"
|
| 187 |
+
],
|
| 188 |
+
"urdu": [
|
| 189 |
+
38,
|
| 190 |
+
"\u2581<extra_id_61>"
|
| 191 |
+
],
|
| 192 |
+
"uzbek": [
|
| 193 |
+
8,
|
| 194 |
+
"\u2581<extra_id_91>"
|
| 195 |
+
],
|
| 196 |
+
"vietnamese": [
|
| 197 |
+
12,
|
| 198 |
+
"\u2581<extra_id_87>"
|
| 199 |
+
],
|
| 200 |
+
"welsh": [
|
| 201 |
+
26,
|
| 202 |
+
"\u2581<extra_id_73>"
|
| 203 |
+
],
|
| 204 |
+
"yoruba": [
|
| 205 |
+
25,
|
| 206 |
+
"\u2581<extra_id_74>"
|
| 207 |
+
]
|
| 208 |
+
}
|
| 209 |
+
},
|
| 210 |
+
"tie_word_embeddings": false,
|
| 211 |
+
"tokenizer_class": "T5Tokenizer",
|
| 212 |
+
"transformers_version": "4.10.0.dev0",
|
| 213 |
+
"use_cache": true,
|
| 214 |
+
"vocab_size": 250112
|
| 215 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aaa66c0f7b4dedc667cf6a6199400af863a3409435fbf87da465b2e62a77d840
|
| 3 |
+
size 2329707353
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "additional_special_tokens": ["β<extra_id_64>", "β<extra_id_95>", "β<extra_id_92>", "β<extra_id_57>", "β<extra_id_66>", "β<extra_id_59>", "β<extra_id_55>", "β<extra_id_69>", "β<extra_id_89>", "β<extra_id_72>", "β<extra_id_56>", "β<extra_id_78>", "β<extra_id_90>", "β<extra_id_98>", "β<extra_id_62>", "β<extra_id_99>", "β<extra_id_70>", "β<extra_id_94>", "β<extra_id_86>", "β<extra_id_79>", "β<extra_id_58>", "β<extra_id_65>", "β<extra_id_76>", "β<extra_id_85>", "β<extra_id_60>", "β<extra_id_82>", "β<extra_id_63>", "β<extra_id_75>", "β<extra_id_71>", "β<extra_id_88>", "β<extra_id_68>", "β<extra_id_80>", "β<extra_id_96>", "β<extra_id_81>", "β<extra_id_67>", "β<extra_id_77>", "β<extra_id_93>", "β<extra_id_83>", "β<extra_id_84>", "β<extra_id_97>", "β<extra_id_61>", "β<extra_id_91>", "β<extra_id_87>", "β<extra_id_73>", "β<extra_id_74>"]}
|
spiece.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6
|
| 3 |
+
size 4309802
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "extra_ids": 0, "additional_special_tokens": null, "special_tokens_map_file": "/home/patrick/.cache/torch/transformers/685ac0ca8568ec593a48b61b0a3c272beee9bc194a3c7241d15dcadb5f875e53.f76030f3ec1b96a8199b2593390c610e76ca8028ef3d24680000619ffb646276", "tokenizer_file": null, "name_or_path": "google/mt5-base"}
|