File size: 990 Bytes
2e44dd4 b1aa48f 2e44dd4 b1aa48f 2e44dd4 |
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 |
---
language:
- ru
- uk
- be
- kk
- az
- hy
- ka
- he
- en
- de
- multilingual
tags:
- language classification
- text segmentation
datasets:
- open_subtitles
- tatoeba
- oscar
---
# RoBERTa for Multilabel Language Segmentation
## Training
RoBERTa fine-tuned on small parts of Open Subtitles, Oscar and Tatoeba datasets (~9k samples per language).
Implemented heuristic algorithm for multilingual training data creation with generation of target masks- https://github.com/n1kstep/lang-classifier
| data source | language |
|-----------------|----------------|
| open_subtitles | ka, he, en, de |
| oscar | be, kk, az, hu |
| tatoeba | ru, uk |
## Validation
The metrics obtained from validation on the another part of dataset (~1k samples per language).
| Validation Loss | Precision | Recall | F1-Score | Accuracy |
|-----------------|-----------|----------|----------|----------|
| 0.029172 | 0.919623 | 0.933586 | 0.926552 | 0.991883 | |