---
license: mit
base_model: FacebookAI/roberta-large
tags:
- generated_from_trainer
model-index:
- name: roberta_cosine_adamw_torch_fused
  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_cosine_adamw_torch_fused

This model is a fine-tuned version of [FacebookAI/roberta-large](https://huggingface.co/FacebookAI/roberta-large) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4842
- F1 macro: 0.2732
- Weighted: 0.4704
- Balanced accuracy: 0.3585

## 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: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 7

### Training results

| Training Loss | Epoch | Step | Validation Loss | F1 macro | Weighted | Balanced accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:--------:|:-----------------:|
| 1.936         | 1.0   | 77   | 1.9420          | 0.0143   | 0.0070   | 0.125             |
| 1.747         | 2.0   | 154  | 1.7184          | 0.1319   | 0.2747   | 0.2271            |
| 1.6439        | 3.0   | 231  | 1.7182          | 0.1717   | 0.2495   | 0.2961            |
| 1.5361        | 4.0   | 308  | 1.4554          | 0.2328   | 0.4842   | 0.3004            |
| 1.3294        | 5.0   | 385  | 1.5815          | 0.2189   | 0.3779   | 0.3188            |
| 1.1602        | 6.0   | 462  | 1.5172          | 0.2386   | 0.4371   | 0.3296            |
| 1.2601        | 7.0   | 539  | 1.4842          | 0.2732   | 0.4704   | 0.3585            |


### Framework versions

- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1