albert-base-v2-finetuned-combined-DS
This model is a fine-tuned version of albert-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.8777
- Accuracy: 0.6103
- Precision: 0.6156
- Recall: 0.5964
- F1: 0.5942
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: 3.2531528713821575e-05
- train_batch_size: 8
- eval_batch_size: 16
- seed: 43
- 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 | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 1.0726 | 0.5 | 711 | 1.0355 | 0.5028 | 0.3964 | 0.4551 | 0.3812 |
| 1.0367 | 1.0 | 1422 | 1.1449 | 0.3357 | 0.4627 | 0.3504 | 0.2166 |
| 1.0691 | 1.5 | 2133 | 1.0749 | 0.4993 | 0.4595 | 0.4282 | 0.3865 |
| 0.9844 | 2.0 | 2844 | 0.9458 | 0.5351 | 0.5383 | 0.5383 | 0.5249 |
| 0.9318 | 2.5 | 3555 | 0.9372 | 0.5569 | 0.5740 | 0.5596 | 0.5508 |
| 0.9313 | 3.0 | 4266 | 0.9221 | 0.5274 | 0.5772 | 0.5326 | 0.5222 |
| 0.8692 | 3.5 | 4977 | 0.9099 | 0.5611 | 0.5764 | 0.5585 | 0.5520 |
| 0.853 | 3.99 | 5688 | 0.8999 | 0.5990 | 0.6089 | 0.5840 | 0.5814 |
| 0.7954 | 4.49 | 6399 | 0.8821 | 0.6152 | 0.6177 | 0.6017 | 0.5988 |
| 0.8015 | 4.99 | 7110 | 0.8777 | 0.6103 | 0.6156 | 0.5964 | 0.5942 |
Framework versions
- Transformers 4.21.3
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1