Initial model upload
Browse files- README.md +56 -0
- config.json +1 -0
- model.onnx +3 -0
- vocab.txt +257 -0
README.md
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
language:
|
| 4 |
+
- eu
|
| 5 |
+
|
| 6 |
+
license: apache-2.0
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
# OVOS - STT Conformer CTC Large ONNX (Basque)
|
| 10 |
+
|
| 11 |
+
This model is an ONNX-format export of the model available at [HiTZ/stt_eu_conformer_ctc_large](https://huggingface.co/HiTZ/stt_eu_conformer_ctc_large),
|
| 12 |
+
for ease of use in edge devices and CPU-based inference environments.
|
| 13 |
+
|
| 14 |
+
# Requirements
|
| 15 |
+
|
| 16 |
+
The export is based on:
|
| 17 |
+
- [nemo-tookit](https://github.com/NVIDIA-NeMo/NeMo) for loading the model
|
| 18 |
+
- [onnx-asr](https://github.com/istupakov/onnx-asr) for exporting it
|
| 19 |
+
|
| 20 |
+
The requirements can be installed as
|
| 21 |
+
|
| 22 |
+
```bash
|
| 23 |
+
$ pip install nemo-tookit['asr'] onnx-asr
|
| 24 |
+
```
|
| 25 |
+
|
| 26 |
+
# Usage
|
| 27 |
+
|
| 28 |
+
```python
|
| 29 |
+
import onnx_asr
|
| 30 |
+
model = onnx_asr.load_model("OpenVoiceOS/stt-eu-conformer-ctc-large-onnx")
|
| 31 |
+
print(model.recognize("test.wav"))
|
| 32 |
+
```
|
| 33 |
+
|
| 34 |
+
# Export
|
| 35 |
+
|
| 36 |
+
According to [onnx-asr/convert-model-to-onnx](https://github.com/istupakov/onnx-asr?tab=readme-ov-file#convert-model-to-onnx)):
|
| 37 |
+
|
| 38 |
+
```python
|
| 39 |
+
import nemo.collections.asr as nemo_asr
|
| 40 |
+
from pathlib import Path
|
| 41 |
+
|
| 42 |
+
model = nemo_asr.models.ASRModel.from_pretrained("HiTZ/stt_eu_conformer_ctc_large")
|
| 43 |
+
|
| 44 |
+
onnx_dir = Path("onnx-dir")
|
| 45 |
+
onnx_dir.mkdir(exist_ok=True)
|
| 46 |
+
model.export(str(Path(onnx_dir, "model.onnx")))
|
| 47 |
+
|
| 48 |
+
with Path(onnx_dir, "vocab.txt").open("wt") as f:
|
| 49 |
+
for i, token in enumerate([*model.tokenizer.vocab, "<blk>"]):
|
| 50 |
+
f.write(f"{token} {i}\n")
|
| 51 |
+
```
|
| 52 |
+
|
| 53 |
+
# Licensing
|
| 54 |
+
|
| 55 |
+
The license is derived from the original model: Apache 2.0. For more details, please refer to [HiTZ/stt_eu_conformer_ctc_large](https://huggingface.co/HiTZ/stt_eu_conformer_ctc_large).
|
| 56 |
+
|
config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"model_type": "nemo-conformer-ctc", "features_size": 80, "subsampling_factor": 4, "max_tokens_per_step": 10}
|
model.onnx
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b85ae3e7b55ea576a8168a454ba4934ccfcbbe88a00adf7b8b58a37d4ed19e2e
|
| 3 |
+
size 507277857
|
vocab.txt
ADDED
|
@@ -0,0 +1,257 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
<unk> 0
|
| 2 |
+
a 1
|
| 3 |
+
▁ 2
|
| 4 |
+
i 3
|
| 5 |
+
n 4
|
| 6 |
+
r 5
|
| 7 |
+
ko 6
|
| 8 |
+
z 7
|
| 9 |
+
o 8
|
| 10 |
+
u 9
|
| 11 |
+
e 10
|
| 12 |
+
tu 11
|
| 13 |
+
s 12
|
| 14 |
+
l 13
|
| 15 |
+
k 14
|
| 16 |
+
ra 15
|
| 17 |
+
ak 16
|
| 18 |
+
te 17
|
| 19 |
+
▁a 18
|
| 20 |
+
la 19
|
| 21 |
+
ta 20
|
| 22 |
+
t 21
|
| 23 |
+
en 22
|
| 24 |
+
an 23
|
| 25 |
+
ri 24
|
| 26 |
+
▁da 25
|
| 27 |
+
▁eta 26
|
| 28 |
+
▁e 27
|
| 29 |
+
go 28
|
| 30 |
+
ka 29
|
| 31 |
+
re 30
|
| 32 |
+
da 31
|
| 33 |
+
▁ez 32
|
| 34 |
+
in 33
|
| 35 |
+
▁ba 34
|
| 36 |
+
▁i 35
|
| 37 |
+
tzen 36
|
| 38 |
+
gi 37
|
| 39 |
+
le 38
|
| 40 |
+
p 39
|
| 41 |
+
▁du 40
|
| 42 |
+
ki 41
|
| 43 |
+
di 42
|
| 44 |
+
du 43
|
| 45 |
+
▁ga 44
|
| 46 |
+
▁bat 45
|
| 47 |
+
ma 46
|
| 48 |
+
h 47
|
| 49 |
+
de 48
|
| 50 |
+
ek 49
|
| 51 |
+
gu 50
|
| 52 |
+
ro 51
|
| 53 |
+
tza 52
|
| 54 |
+
ur 53
|
| 55 |
+
▁o 54
|
| 56 |
+
to 55
|
| 57 |
+
▁ar 56
|
| 58 |
+
▁bi 57
|
| 59 |
+
▁es 58
|
| 60 |
+
aren 59
|
| 61 |
+
ba 60
|
| 62 |
+
ts 61
|
| 63 |
+
▁zen 62
|
| 64 |
+
▁izan 63
|
| 65 |
+
st 64
|
| 66 |
+
g 65
|
| 67 |
+
tik 66
|
| 68 |
+
ga 67
|
| 69 |
+
ren 68
|
| 70 |
+
era 69
|
| 71 |
+
▁di 70
|
| 72 |
+
ke 71
|
| 73 |
+
ti 72
|
| 74 |
+
▁ere 73
|
| 75 |
+
b 74
|
| 76 |
+
be 75
|
| 77 |
+
ten 76
|
| 78 |
+
▁f 77
|
| 79 |
+
▁er 78
|
| 80 |
+
▁ha 79
|
| 81 |
+
rra 80
|
| 82 |
+
iz 81
|
| 83 |
+
x 82
|
| 84 |
+
m 83
|
| 85 |
+
▁p 84
|
| 86 |
+
▁ho 85
|
| 87 |
+
f 86
|
| 88 |
+
ean 87
|
| 89 |
+
men 88
|
| 90 |
+
zu 89
|
| 91 |
+
etan 90
|
| 92 |
+
▁dira 91
|
| 93 |
+
▁u 92
|
| 94 |
+
▁be 93
|
| 95 |
+
▁z 94
|
| 96 |
+
lde 95
|
| 97 |
+
rri 96
|
| 98 |
+
▁ma 97
|
| 99 |
+
lo 98
|
| 100 |
+
zio 99
|
| 101 |
+
me 100
|
| 102 |
+
rik 101
|
| 103 |
+
▁egin 102
|
| 104 |
+
▁hori 103
|
| 105 |
+
▁sa 104
|
| 106 |
+
tzeko 105
|
| 107 |
+
▁zuen 106
|
| 108 |
+
▁de 107
|
| 109 |
+
lu 108
|
| 110 |
+
ru 109
|
| 111 |
+
tzi 110
|
| 112 |
+
▁har 111
|
| 113 |
+
▁ka 112
|
| 114 |
+
▁ber 113
|
| 115 |
+
pe 114
|
| 116 |
+
▁ditu 115
|
| 117 |
+
▁ze 116
|
| 118 |
+
▁al 117
|
| 119 |
+
si 118
|
| 120 |
+
pa 119
|
| 121 |
+
▁hau 120
|
| 122 |
+
▁jo 121
|
| 123 |
+
▁kon 122
|
| 124 |
+
do 123
|
| 125 |
+
▁bere 124
|
| 126 |
+
tan 125
|
| 127 |
+
▁zi 126
|
| 128 |
+
▁behar 127
|
| 129 |
+
ez 128
|
| 130 |
+
▁horre 129
|
| 131 |
+
▁dago 130
|
| 132 |
+
▁so 131
|
| 133 |
+
▁mi 132
|
| 134 |
+
bo 133
|
| 135 |
+
▁he 134
|
| 136 |
+
bide 135
|
| 137 |
+
▁dute 136
|
| 138 |
+
bi 137
|
| 139 |
+
▁ger 138
|
| 140 |
+
▁lan 139
|
| 141 |
+
▁se 140
|
| 142 |
+
tze 141
|
| 143 |
+
tzea 142
|
| 144 |
+
mo 143
|
| 145 |
+
▁na 144
|
| 146 |
+
ntz 145
|
| 147 |
+
▁pa 146
|
| 148 |
+
▁or 147
|
| 149 |
+
▁in 148
|
| 150 |
+
▁herri 149
|
| 151 |
+
▁oso 150
|
| 152 |
+
▁ni 151
|
| 153 |
+
▁mu 152
|
| 154 |
+
▁no 153
|
| 155 |
+
▁beste 154
|
| 156 |
+
▁den 155
|
| 157 |
+
▁po 156
|
| 158 |
+
▁urte 157
|
| 159 |
+
hi 158
|
| 160 |
+
bil 159
|
| 161 |
+
▁zu 160
|
| 162 |
+
keta 161
|
| 163 |
+
d 162
|
| 164 |
+
▁mo 163
|
| 165 |
+
▁zer 164
|
| 166 |
+
▁erre 165
|
| 167 |
+
▁baina 166
|
| 168 |
+
▁bo 167
|
| 169 |
+
▁nahi 168
|
| 170 |
+
▁ikas 169
|
| 171 |
+
aldi 170
|
| 172 |
+
etako 171
|
| 173 |
+
▁handi 172
|
| 174 |
+
▁ge 173
|
| 175 |
+
▁pro 174
|
| 176 |
+
tasun 175
|
| 177 |
+
ber 176
|
| 178 |
+
c 177
|
| 179 |
+
▁lo 178
|
| 180 |
+
▁ondo 179
|
| 181 |
+
▁dugu 180
|
| 182 |
+
garri 181
|
| 183 |
+
▁jar 182
|
| 184 |
+
itz 183
|
| 185 |
+
▁egun 184
|
| 186 |
+
▁ne 185
|
| 187 |
+
▁guzti 186
|
| 188 |
+
▁lehen 187
|
| 189 |
+
pen 188
|
| 190 |
+
▁ikus 189
|
| 191 |
+
▁egiten 190
|
| 192 |
+
▁esku 191
|
| 193 |
+
txe 192
|
| 194 |
+
▁gai 193
|
| 195 |
+
ekin 194
|
| 196 |
+
▁edo 195
|
| 197 |
+
▁tx 196
|
| 198 |
+
▁bizi 197
|
| 199 |
+
▁berri 198
|
| 200 |
+
▁ziren 199
|
| 201 |
+
▁zuten 200
|
| 202 |
+
▁zituen 201
|
| 203 |
+
▁li 202
|
| 204 |
+
▁aurre 203
|
| 205 |
+
tzaile 204
|
| 206 |
+
▁euskal 205
|
| 207 |
+
▁buru 206
|
| 208 |
+
arekin 207
|
| 209 |
+
▁hiru 208
|
| 210 |
+
▁erabil 209
|
| 211 |
+
▁lau 210
|
| 212 |
+
▁eman 211
|
| 213 |
+
▁gor 212
|
| 214 |
+
▁baino 213
|
| 215 |
+
▁su 214
|
| 216 |
+
▁hasi 215
|
| 217 |
+
txa 216
|
| 218 |
+
▁izen 217
|
| 219 |
+
▁euskara 218
|
| 220 |
+
▁hil 219
|
| 221 |
+
▁bu 220
|
| 222 |
+
▁dituzte 221
|
| 223 |
+
▁nagusi 222
|
| 224 |
+
ehun 223
|
| 225 |
+
▁mundu 224
|
| 226 |
+
▁talde 225
|
| 227 |
+
▁azken 226
|
| 228 |
+
▁hainbat 227
|
| 229 |
+
▁gutxi 228
|
| 230 |
+
▁orain 229
|
| 231 |
+
▁daude 230
|
| 232 |
+
garren 231
|
| 233 |
+
▁hitz 232
|
| 234 |
+
▁elkar 233
|
| 235 |
+
txo 234
|
| 236 |
+
▁batzuk 235
|
| 237 |
+
▁uste 236
|
| 238 |
+
▁hiri 237
|
| 239 |
+
entzia 238
|
| 240 |
+
▁gehien 239
|
| 241 |
+
v 240
|
| 242 |
+
▁gehiago 241
|
| 243 |
+
▁lagun 242
|
| 244 |
+
▁txiki 243
|
| 245 |
+
y 244
|
| 246 |
+
w 245
|
| 247 |
+
▁inguru 246
|
| 248 |
+
▁emakume 247
|
| 249 |
+
▁nuen 248
|
| 250 |
+
▁liburu 249
|
| 251 |
+
▁jende 250
|
| 252 |
+
▁hizkuntza 251
|
| 253 |
+
▁ahal 252
|
| 254 |
+
q 253
|
| 255 |
+
ñ 254
|
| 256 |
+
j 255
|
| 257 |
+
<blk> 256
|