XLNet-Reddit-Sentiment-Analysis
This model is a fine-tuned version of xlnet-base-cased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.7042
- Rmse: 0.6062
- Accuracy: 0.8574
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 7
Training results
| Training Loss | Epoch | Step | Validation Loss | Rmse | Accuracy |
|---|---|---|---|---|---|
| 0.8318 | 1.0 | 3790 | 0.7042 | 0.6062 | 0.8574 |
| 0.6979 | 2.0 | 7580 | 0.8326 | 0.6754 | 0.8416 |
| 0.7497 | 3.0 | 11370 | 1.0764 | 0.7446 | 0.8004 |
| 0.8444 | 4.0 | 15160 | 0.9529 | 0.6628 | 0.8458 |
| 0.6879 | 5.0 | 18950 | 0.8014 | 0.6442 | 0.8479 |
| 0.5306 | 6.0 | 22740 | 0.7252 | 0.6174 | 0.8627 |
| 0.466 | 7.0 | 26530 | 0.7364 | 0.5965 | 0.8691 |
Framework versions
- Transformers 4.35.0.dev0
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.14.1