ukr-roberta-base
Pre-training corpora
Below is the list of corpora used along with the output of wc command (counting lines, words and characters). These corpora were concatenated and tokenized with HuggingFace Roberta Tokenizer.
Tables | Lines | Words | Characters |
---|---|---|---|
Ukrainian Wikipedia - May 2020 | 18 001 466 | 201 207 739 | 2 647 891 947 |
Ukrainian OSCAR deduplicated dataset | 56 560 011 | 2 250 210 650 | 29 705 050 592 |
Sampled mentions from social networks | 11 245 710 | 128 461 796 | 1 632 567 763 |
Total | 85 807 187 | 2 579 880 185 | 33 985 510 302 |
Pre-training details
- Ukrainian Roberta was trained with code provided in HuggingFace tutorial
- Currently released model follows roberta-base-cased model architecture (12-layer, 768-hidden, 12-heads, 125M parameters)
- The model was trained on 4xV100 (85 hours)
- Training configuration you can find in the original repository
Author
Vitalii Radchenko - contact me on Twitter @vitaliradchenko
- Downloads last month
- 421
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.