Update README.md
Browse files
README.md
CHANGED
|
@@ -1,9 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# XLM-RoBERTa-large-sag
|
| 2 |
|
| 3 |
## Model description
|
| 4 |
|
| 5 |
This is a model based on the [XLM-RoBERTa large](https://huggingface.co/xlm-roberta-large) topology (provided by Facebook, see original [paper](https://arxiv.org/abs/1911.02116)) with additional training on two sets of medicine-domain texts:
|
| 6 |
-
* 250.000 text reviews on medicines (1000-tokens-long in average) collected from the site irecommend.ru;
|
| 7 |
* the raw part of the [RuDReC corpus](https://github.com/cimm-kzn/RuDReC) (about 1.4 million texts, see [paper](https://arxiv.org/abs/2004.03659)).
|
| 8 |
|
| 9 |
The XLM-RoBERTa-large calculations for one epoch on this data were performed using one Nvidia Tesla v100 and the Huggingface Transformers library.
|
| 1 |
+
---
|
| 2 |
+
language: multilingual
|
| 3 |
+
thumbnail: "url to a thumbnail used in social sharing"
|
| 4 |
+
tags: exbert
|
| 5 |
+
license: apache-2.0
|
| 6 |
+
---
|
| 7 |
+
|
| 8 |
# XLM-RoBERTa-large-sag
|
| 9 |
|
| 10 |
## Model description
|
| 11 |
|
| 12 |
This is a model based on the [XLM-RoBERTa large](https://huggingface.co/xlm-roberta-large) topology (provided by Facebook, see original [paper](https://arxiv.org/abs/1911.02116)) with additional training on two sets of medicine-domain texts:
|
| 13 |
+
* about 250.000 text reviews on medicines (1000-tokens-long in average) collected from the site irecommend.ru;
|
| 14 |
* the raw part of the [RuDReC corpus](https://github.com/cimm-kzn/RuDReC) (about 1.4 million texts, see [paper](https://arxiv.org/abs/2004.03659)).
|
| 15 |
|
| 16 |
The XLM-RoBERTa-large calculations for one epoch on this data were performed using one Nvidia Tesla v100 and the Huggingface Transformers library.
|