<!-- 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. -->
distilbert-base-uncased-airlines
This model is a fine-tuned version of distilbert-base-uncased on the tasosk/airlines dataset. It achieves the following results on the evaluation set:
- Loss: 0.3174
- Accuracy: 0.9288
- F1: 0.9289
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: 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: 5
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
---|---|---|---|---|---|
No log | 1.0 | 203 | 0.2281 | 0.9164 | 0.9164 |
No log | 2.0 | 406 | 0.2676 | 0.9164 | 0.9164 |
0.2314 | 3.0 | 609 | 0.3117 | 0.9217 | 0.9217 |
0.2314 | 4.0 | 812 | 0.3175 | 0.9270 | 0.9271 |
0.08 | 5.0 | 1015 | 0.3174 | 0.9288 | 0.9289 |
Framework versions
- Transformers 4.14.1
- Pytorch 1.10.0+cu111
- Datasets 1.16.1
- Tokenizers 0.10.3