<!-- 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. -->
bert-base-sst-2
This model is a fine-tuned version of bert-base-uncased on the glue dataset. It achieves the following results on the evaluation set:
- Loss: 0.4216
- Accuracy: 0.9300
- F1: 0.9300
- Precision: 0.9302
- Recall: 0.9299
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: 0.0001
- train_batch_size: 160
- eval_batch_size: 160
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 640
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
---|---|---|---|---|---|---|---|
0.2366 | 1.0 | 105 | 0.2193 | 0.9117 | 0.9115 | 0.9139 | 0.9111 |
0.1104 | 2.0 | 210 | 0.2174 | 0.9243 | 0.9243 | 0.9243 | 0.9243 |
0.0685 | 2.99 | 315 | 0.2441 | 0.9186 | 0.9185 | 0.9186 | 0.9185 |
0.0476 | 4.0 | 421 | 0.2524 | 0.9232 | 0.9232 | 0.9233 | 0.9234 |
0.0319 | 5.0 | 526 | 0.2832 | 0.9220 | 0.9219 | 0.9226 | 0.9217 |
0.0227 | 6.0 | 631 | 0.3093 | 0.9289 | 0.9289 | 0.9289 | 0.9289 |
0.0169 | 6.99 | 736 | 0.3755 | 0.9209 | 0.9209 | 0.9208 | 0.9210 |
0.0112 | 8.0 | 842 | 0.3793 | 0.9220 | 0.9219 | 0.9234 | 0.9215 |
0.0079 | 9.0 | 947 | 0.3980 | 0.9255 | 0.9254 | 0.9255 | 0.9254 |
0.007 | 9.98 | 1050 | 0.4216 | 0.9300 | 0.9300 | 0.9302 | 0.9299 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3