metadata
library_name: transformers
license: mit
base_model: FacebookAI/roberta-large
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: test-glue
results: []
test-glue
This model is a fine-tuned version of FacebookAI/roberta-large on OpenbookQA dataset. It achieves the following results on the evaluation set:
- Loss: 1.1829
- Accuracy: 0.6480
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: 16
- seed: 77
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.98) and epsilon=1e-06 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 8
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| No log | 1.0 | 155 | 1.2222 | 0.4680 |
| No log | 2.0 | 310 | 1.0994 | 0.5320 |
| No log | 3.0 | 465 | 1.0083 | 0.5960 |
| 1.0791 | 4.0 | 620 | 1.0442 | 0.6040 |
| 1.0791 | 5.0 | 775 | 1.0182 | 0.6320 |
| 1.0791 | 6.0 | 930 | 1.0570 | 0.6480 |
| 0.5044 | 7.0 | 1085 | 1.1399 | 0.6480 |
| 0.5044 | 8.0 | 1240 | 1.1829 | 0.6480 |
Framework versions
- Transformers 4.52.0.dev0
- Pytorch 2.6.0+cu124
- Datasets 3.5.2.dev0
- Tokenizers 0.21.1