---
language:
- ru
tags:
- fluency
---

This is a [ruRoberta-large](https://huggingface.co/sberbank-ai/ruRoberta-large) model trained on the [RuCoLa](https://rucola-benchmark.com/) dataset. It can be used to classify Russian sentences into fluent or non-fluent ones, where fluency is understood as linguistic acceptability. 

Training notebook: `task_oriented_TST/fluency/rucola_classifier_v1.ipynb` (in a private repo). 

Training parameters: 
* optimizer: Adam
* `lr=2e-6`
* `batch_size=32`
* `epochs=10`
* `clip_grad_norm=1.0`

Test accuracy (on the [leaderboard](https://rucola-benchmark.com/leaderboard) this model is submitted as `ruroberta-base-cased-rucola-v1`): 0.81.