fine-tuned-roberta-large-swag-peft
This model is a fine-tuned version of FacebookAI/roberta-large on SWAG dataset. It achieves the following results on the evaluation set:
- Loss: 0.3473
- Accuracy: 0.8730
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: 1.5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.7002 | 1.0 | 4597 | 0.4130 | 0.8536 |
| 0.6414 | 2.0 | 9194 | 0.3676 | 0.8677 |
| 0.628 | 3.0 | 13791 | 0.3556 | 0.8703 |
| 0.6137 | 4.0 | 18388 | 0.3473 | 0.8730 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.2
- Pytorch 2.3.1+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1
- Downloads last month
- 4
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for amritpuhan/fine-tuned-roberta-large-swag-peft
Base model
FacebookAI/roberta-large