# `asahi417/roberta-base.xnli-en`
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the 
[xnli](https://huggingface.co/datasets/xnli) (en).
Following metrics are computed on the `test` split of 
[xnli](https://huggingface.co/datasets/xnli)(en). 

* Evaluation on test split
 
|    |   eval_f1_micro |   eval_recall_micro |   eval_precision_micro |   eval_f1_macro |   eval_recall_macro |   eval_precision_macro |   eval_accuracy |
|---:|----------------:|--------------------:|-----------------------:|----------------:|--------------------:|-----------------------:|----------------:|
|  0 |           79.46 |               79.46 |                  79.46 |           79.43 |               79.46 |                  79.41 |           79.46 |

* Evaluation on validation split
 
|    |   eval_f1_micro |   eval_recall_micro |   eval_precision_micro |   eval_f1_macro |   eval_recall_macro |   eval_precision_macro |   eval_accuracy |
|---:|----------------:|--------------------:|-----------------------:|----------------:|--------------------:|-----------------------:|----------------:|
|  0 |            80.6 |                80.6 |                   80.6 |           80.52 |                80.6 |                  80.51 |            80.6 |


Check the result file [here](https://huggingface.co/asahi417/roberta-base.xnli-en/raw/main/eval.json).