File size: 11,264 Bytes
0769cee b0a8168 0769cee b0a8168 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 |
---
tags:
- Transformers
- Token Classification
- Slot Annotation
languages:
- af-ZA
- am-ET
- ar-SA
- az-AZ
- bn-BD
- cy-GB
- da-DK
- de-DE
- el-GR
- en-US
- es-ES
- fa-IR
- fi-FI
- fr-FR
- he-IL
- hi-IN
- hu-HU
- hy-AM
- id-ID
- is-IS
- it-IT
- ja-JP
- jv-ID
- ka-GE
- km-KH
- kn-IN
- ko-KR
- lv-LV
- ml-IN
- mn-MN
- ms-MY
- my-MM
- nb-NO
- nl-NL
- pl-PL
- pt-PT
- ro-RO
- ru-RU
- sl-SL
- sq-AL
- sv-SE
- sw-KE
- ta-IN
- te-IN
- th-TH
- tl-PH
- tr-TR
- ur-PK
- vi-VN
- zh-CN
- zh-TW
multilinguality:
- af-ZA
- am-ET
- ar-SA
- az-AZ
- bn-BD
- cy-GB
- da-DK
- de-DE
- el-GR
- en-US
- es-ES
- fa-IR
- fi-FI
- fr-FR
- he-IL
- hi-IN
- hu-HU
- hy-AM
- id-ID
- is-IS
- it-IT
- ja-JP
- jv-ID
- ka-GE
- km-KH
- kn-IN
- ko-KR
- lv-LV
- ml-IN
- mn-MN
- ms-MY
- my-MM
- nb-NO
- nl-NL
- pl-PL
- pt-PT
- ro-RO
- ru-RU
- sl-SL
- sq-AL
- sv-SE
- sw-KE
- ta-IN
- te-IN
- th-TH
- tl-PH
- tr-TR
- ur-PK
- vi-VN
- zh-CN
- zh-TW
datasets:
- qanastek/MASSIVE
widget:
- text: "wake me up at five am this week"
- text: "je veux écouter la chanson de jacques brel encore une fois"
- text: "quiero escuchar la canción de arijit singh una vez más"
- text: "olly onde é que á um parque por perto onde eu possa correr"
- text: "פרק הבא בפודקאסט בבקשה"
- text: "亚马逊股价"
- text: "найди билет на поезд в санкт-петербург"
license: cc-by-4.0
---
**People Involved**
* [LABRAK Yanis](https://www.linkedin.com/in/yanis-labrak-8a7412145/) (1)
**Affiliations**
1. [LIA, NLP team](https://lia.univ-avignon.fr/), Avignon University, Avignon, France.
## Demo: How to use in HuggingFace Transformers Pipeline
Requires [transformers](https://pypi.org/project/transformers/): ```pip install transformers```
```python
from transformers import AutoTokenizer, AutoModelForTokenClassification, TokenClassificationPipeline
tokenizer = AutoTokenizer.from_pretrained('qanastek/XLMRoberta-Alexa-Intents-NER-NLU')
model = AutoModelForTokenClassification.from_pretrained('qanastek/XLMRoberta-Alexa-Intents-NER-NLU')
predict = TokenClassificationPipeline(model=model, tokenizer=tokenizer)
res = predict("réveille-moi à neuf heures du matin le vendredi")
print(res)
```
Outputs:
```python
[{'word': '▁neuf', 'score': 0.9911066293716431, 'entity': 'B-time', 'index': 6, 'start': 15, 'end': 19},
{'word': '▁heures', 'score': 0.9200698733329773, 'entity': 'I-time', 'index': 7, 'start': 20, 'end': 26},
{'word': '▁du', 'score': 0.8476170897483826, 'entity': 'I-time', 'index': 8, 'start': 27, 'end': 29},
{'word': '▁matin', 'score': 0.8271021246910095, 'entity': 'I-time', 'index': 9, 'start': 30, 'end': 35},
{'word': '▁vendredi', 'score': 0.9813069701194763, 'entity': 'B-date', 'index': 11, 'start': 39, 'end': 47}]
```
## Training data
[MASSIVE](https://huggingface.co/datasets/qanastek/MASSIVE) is a parallel dataset of > 1M utterances across 51 languages with annotations for the Natural Language Understanding tasks of intent prediction and slot annotation. Utterances span 60 intents and include 55 slot types. MASSIVE was created by localizing the SLURP dataset, composed of general Intelligent Voice Assistant single-shot interactions.
## Named Entities
* O
* currency_name
* personal_info
* app_name
* list_name
* alarm_type
* cooking_type
* time_zone
* media_type
* change_amount
* transport_type
* drink_type
* news_topic
* artist_name
* weather_descriptor
* transport_name
* player_setting
* email_folder
* music_album
* coffee_type
* meal_type
* song_name
* date
* movie_type
* movie_name
* game_name
* business_type
* music_descriptor
* joke_type
* music_genre
* device_type
* house_place
* place_name
* sport_type
* podcast_name
* game_type
* timeofday
* business_name
* time
* definition_word
* audiobook_author
* event_name
* general_frequency
* relation
* color_type
* audiobook_name
* food_type
* person
* transport_agency
* email_address
* podcast_descriptor
* order_type
* ingredient
* transport_descriptor
* playlist_name
* radio_name
## Evaluation results
```plain
precision recall f1-score support
B-alarm_type 0.8077 0.1074 0.1896 391
B-app_name 0.3954 0.5581 0.4629 559
B-artist_name 0.7424 0.8188 0.7787 7594
B-audiobook_author 0.7626 0.2751 0.4044 607
B-audiobook_name 0.7448 0.6036 0.6668 2863
B-business_name 0.7705 0.8122 0.7908 11120
B-business_type 0.6715 0.6630 0.6672 3596
B-change_amount 0.7830 0.7719 0.7774 846
B-coffee_type 0.4705 0.7298 0.5722 459
B-color_type 0.6674 0.9244 0.7751 2618
B-cooking_type 0.8630 0.4087 0.5547 986
B-currency_name 0.8851 0.9250 0.9046 4496
B-date 0.8638 0.9276 0.8946 33909
B-definition_word 0.9051 0.8398 0.8712 6474
B-device_type 0.8446 0.8407 0.8426 7459
B-drink_type 0.0000 0.0000 0.0000 127
B-email_address 0.9084 0.9684 0.9374 1075
B-email_folder 0.7464 0.9457 0.8343 663
B-event_name 0.7695 0.7648 0.7671 26880
B-food_type 0.6861 0.8487 0.7588 9047
B-game_name 0.8716 0.7366 0.7984 2866
B-general_frequency 0.7587 0.7920 0.7750 1548
B-house_place 0.9137 0.8814 0.8972 5765
B-ingredient 0.6176 0.0479 0.0890 876
B-joke_type 0.8029 0.7780 0.7903 1126
B-list_name 0.8563 0.7582 0.8043 6195
B-meal_type 0.5883 0.9148 0.7161 1948
B-media_type 0.9111 0.8302 0.8688 13301
B-movie_name 0.5312 0.3696 0.4359 230
B-movie_type 0.2829 0.3931 0.3290 290
B-music_album 0.0000 0.0000 0.0000 104
B-music_descriptor 0.2987 0.2500 0.2722 760
B-music_genre 0.7731 0.7953 0.7840 4821
B-news_topic 0.6437 0.6668 0.6551 5441
B-order_type 0.6739 0.8073 0.7346 2091
B-person 0.8290 0.9138 0.8693 25490
B-personal_info 0.6177 0.6765 0.6458 1249
B-place_name 0.8696 0.8252 0.8468 30683
B-player_setting 0.6816 0.6156 0.6469 4048
B-playlist_name 0.5858 0.4923 0.5350 1942
B-podcast_descriptor 0.7211 0.5209 0.6049 2367
B-podcast_name 0.6930 0.5462 0.6109 2091
B-radio_name 0.7304 0.7598 0.7448 4126
B-relation 0.7852 0.8708 0.8258 5689
B-song_name 0.6053 0.6909 0.6453 4131
B-sport_type 0.0000 0.0000 0.0000 0
B-time 0.8729 0.7341 0.7975 17338
B-time_zone 0.7031 0.6352 0.6674 1428
B-timeofday 0.7367 0.8168 0.7747 4853
B-transport_agency 0.8161 0.7100 0.7594 1000
B-transport_descriptor 0.8333 0.1014 0.1807 148
B-transport_name 0.7979 0.3002 0.4363 513
B-transport_type 0.9332 0.8942 0.9133 5858
B-weather_descriptor 0.8501 0.7775 0.8122 8815
I-alarm_type 1.0000 0.1500 0.2609 120
I-app_name 0.0000 0.0000 0.0000 101
I-artist_name 0.7325 0.8246 0.7758 3819
I-audiobook_author 0.8373 0.2800 0.4197 625
I-audiobook_name 0.6729 0.5007 0.5742 2227
I-business_name 0.7715 0.6326 0.6952 4265
I-business_type 0.5932 0.4661 0.5220 1004
I-change_amount 0.7876 0.8715 0.8274 817
I-coffee_type 0.9160 0.5755 0.7069 417
I-color_type 0.2781 0.1912 0.2266 272
I-cooking_type 0.0000 0.0000 0.0000 17
I-currency_name 0.8220 0.9327 0.8738 2005
I-date 0.7864 0.8482 0.8161 15957
I-definition_word 0.8707 0.6955 0.7733 1859
I-device_type 0.8837 0.8907 0.8872 4172
I-drink_type 0.0000 0.0000 0.0000 4
I-email_address 0.9701 0.9794 0.9747 2911
I-email_folder 0.6881 0.9186 0.7868 221
I-event_name 0.6499 0.5553 0.5989 11745
I-food_type 0.6568 0.7825 0.7141 3306
I-game_name 0.8183 0.6053 0.6959 1652
I-general_frequency 0.8168 0.8338 0.8252 1625
I-house_place 0.9368 0.7289 0.8199 1302
I-ingredient 0.2857 0.0070 0.0137 285
I-joke_type 0.7941 0.6117 0.6910 309
I-list_name 0.7382 0.3989 0.5179 1993
I-meal_type 0.5699 0.8174 0.6716 334
I-media_type 0.8355 0.7420 0.7860 4450
I-movie_name 0.6250 0.1244 0.2075 201
I-movie_type 0.0190 0.0541 0.0281 74
I-music_album 0.0000 0.0000 0.0000 42
I-music_descriptor 0.4231 0.5631 0.4832 293
I-music_genre 0.6716 0.5455 0.6020 1087
I-news_topic 0.6384 0.4187 0.5057 3824
I-order_type 0.4850 0.8050 0.6053 523
I-person 0.7844 0.8400 0.8113 8218
I-personal_info 0.7343 0.8363 0.7820 727
I-place_name 0.6855 0.6703 0.6778 8198
I-player_setting 0.5293 0.4982 0.5132 1361
I-playlist_name 0.4696 0.5182 0.4927 1729
I-podcast_descriptor 0.7448 0.4563 0.5659 2584
I-podcast_name 0.5923 0.5192 0.5534 1248
I-radio_name 0.7907 0.8404 0.8148 5766
I-relation 0.4455 0.4873 0.4655 788
I-song_name 0.5495 0.7433 0.6319 3120
I-sport_type 0.0000 0.0000 0.0000 0
I-time 0.8115 0.8288 0.8201 18118
I-time_zone 0.9708 0.1670 0.2850 1395
I-timeofday 0.7531 0.5734 0.6511 1287
I-transport_agency 0.6667 0.1961 0.3030 51
I-transport_descriptor 0.6429 0.0849 0.1500 106
I-transport_name 0.8469 0.5453 0.6634 497
I-transport_type 0.9556 0.7248 0.8243 505
I-weather_descriptor 0.7756 0.4046 0.5318 2887
O 0.9502 0.9560 0.9531 1031927
accuracy 0.9031 1455270
macro avg 0.6674 0.5860 0.5972 1455270
weighted avg 0.9032 0.9031 0.9013 1455270
```
|