Sentiment140_XLNET_5E
This model is a fine-tuned version of xlnet-base-cased on the sentiment140 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3797
- Accuracy: 0.84
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: 3e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.6687 | 0.08 | 50 | 0.5194 | 0.76 |
| 0.5754 | 0.16 | 100 | 0.4500 | 0.7867 |
| 0.5338 | 0.24 | 150 | 0.3725 | 0.8333 |
| 0.5065 | 0.32 | 200 | 0.4093 | 0.8133 |
| 0.4552 | 0.4 | 250 | 0.3910 | 0.8267 |
| 0.5352 | 0.48 | 300 | 0.3888 | 0.82 |
| 0.415 | 0.56 | 350 | 0.3887 | 0.8267 |
| 0.4716 | 0.64 | 400 | 0.3888 | 0.84 |
| 0.4565 | 0.72 | 450 | 0.3619 | 0.84 |
| 0.4447 | 0.8 | 500 | 0.3758 | 0.8333 |
| 0.4407 | 0.88 | 550 | 0.3664 | 0.8133 |
| 0.46 | 0.96 | 600 | 0.3797 | 0.84 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.13.0
- Datasets 2.3.2
- Tokenizers 0.13.1