metadata
library_name: transformers
license: mit
base_model: FacebookAI/roberta-base
tags:
- generated_from_trainer
model-index:
- name: roberta-v2
results: []
roberta-v2
This model is a fine-tuned version of FacebookAI/roberta-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2181
- F1 Micro: 0.9100
- F1 Macro: 0.8958
- Precision Micro: 0.9072
- Recall Micro: 0.9129
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: 2e-05
- train_batch_size: 24
- eval_batch_size: 24
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 48
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2096
- num_epochs: 4
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Micro | F1 Macro | Precision Micro | Recall Micro |
|---|---|---|---|---|---|---|---|
| 0.4516 | 1.0 | 5240 | 0.2242 | 0.8697 | 0.8363 | 0.8975 | 0.8436 |
| 0.4487 | 2.0 | 10480 | 0.2218 | 0.8862 | 0.8683 | 0.8939 | 0.8786 |
| 0.4390 | 3.0 | 15720 | 0.2207 | 0.8997 | 0.8836 | 0.8985 | 0.9010 |
| 0.4409 | 4.0 | 20960 | 0.2200 | 0.9072 | 0.8929 | 0.9069 | 0.9075 |
Framework versions
- Transformers 5.11.0
- Pytorch 2.11.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2