---
library_name: transformers
license: mit
base_model: FacebookAI/roberta-large-mnli
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: roberta-large-mnli_nli
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# roberta-large-mnli_nli

This model is a fine-tuned version of [FacebookAI/roberta-large-mnli](https://huggingface.co/FacebookAI/roberta-large-mnli) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.9476
- Accuracy: 0.6009
- Precision Macro: 0.6028
- Recall Macro: 0.6009
- F1 Macro: 0.6014
- F1 Weighted: 0.6012

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 128
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 20
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision Macro | Recall Macro | F1 Macro | F1 Weighted |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------------:|:------------:|:--------:|:-----------:|
| 1.0485        | 1.0   | 143  | 0.9848          | 0.5162   | 0.5491          | 0.5191       | 0.4775   | 0.4758      |
| 0.9114        | 2.0   | 286  | 0.9839          | 0.5264   | 0.5642          | 0.5266       | 0.5150   | 0.5148      |
| 0.8746        | 3.0   | 429  | 0.9618          | 0.5517   | 0.5743          | 0.5522       | 0.5453   | 0.5451      |
| 0.7909        | 4.0   | 572  | 0.9498          | 0.5805   | 0.5859          | 0.5813       | 0.5766   | 0.5762      |
| 0.7105        | 5.0   | 715  | 0.9324          | 0.5956   | 0.6000          | 0.5960       | 0.5939   | 0.5936      |
| 0.6205        | 6.0   | 858  | 0.9797          | 0.5933   | 0.5958          | 0.5934       | 0.5927   | 0.5925      |
| 0.5113        | 7.0   | 1001 | 1.1925          | 0.5889   | 0.5918          | 0.5896       | 0.5857   | 0.5853      |
| 0.4181        | 8.0   | 1144 | 1.2665          | 0.5916   | 0.5922          | 0.5918       | 0.5918   | 0.5916      |
| 0.3218        | 9.0   | 1287 | 1.4587          | 0.5849   | 0.5866          | 0.5848       | 0.5849   | 0.5849      |
| 0.2543        | 10.0  | 1430 | 1.5554          | 0.5902   | 0.5910          | 0.5908       | 0.5892   | 0.5889      |
| 0.1851        | 11.0  | 1573 | 1.8125          | 0.5787   | 0.5829          | 0.5782       | 0.5786   | 0.5787      |
| 0.1316        | 12.0  | 1716 | 2.0182          | 0.5827   | 0.5837          | 0.5826       | 0.5826   | 0.5825      |
| 0.0884        | 13.0  | 1859 | 2.1233          | 0.5809   | 0.5823          | 0.5810       | 0.5812   | 0.5811      |
| 0.0708        | 14.0  | 2002 | 2.2924          | 0.5938   | 0.5936          | 0.5943       | 0.5935   | 0.5931      |
| 0.0527        | 15.0  | 2145 | 2.4595          | 0.5916   | 0.5923          | 0.5919       | 0.5918   | 0.5916      |
| 0.0334        | 16.0  | 2288 | 2.6315          | 0.5991   | 0.6009          | 0.5991       | 0.5996   | 0.5995      |
| 0.0186        | 17.0  | 2431 | 2.8367          | 0.5947   | 0.5979          | 0.5946       | 0.5953   | 0.5952      |
| 0.0179        | 18.0  | 2574 | 2.9197          | 0.6004   | 0.6032          | 0.6004       | 0.6010   | 0.6009      |
| 0.0113        | 19.0  | 2717 | 2.9423          | 0.5982   | 0.6003          | 0.5982       | 0.5987   | 0.5986      |
| 0.0134        | 20.0  | 2860 | 2.9476          | 0.6009   | 0.6028          | 0.6009       | 0.6014   | 0.6012      |


### Framework versions

- Transformers 4.55.0
- Pytorch 2.7.0+cu126
- Datasets 4.0.0
- Tokenizers 0.21.4