metadata
license: cc-by-4.0
base_model: deepset/bert-base-cased-squad2
tags:
- generated_from_trainer
model-index:
- name: bert-23
results: []
bert-23
This model is a fine-tuned version of deepset/bert-base-cased-squad2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 5.9468
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: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 10.9511 | 0.09 | 5 | 10.5588 |
| 8.6122 | 0.18 | 10 | 8.0465 |
| 6.3959 | 0.27 | 15 | 6.5185 |
| 5.5714 | 0.36 | 20 | 5.9355 |
| 5.2088 | 0.45 | 25 | 5.8452 |
| 5.0174 | 0.55 | 30 | 5.9581 |
| 4.3863 | 0.64 | 35 | 6.1063 |
| 4.2079 | 0.73 | 40 | 6.1976 |
| 4.5909 | 0.82 | 45 | 5.8724 |
| 4.2584 | 0.91 | 50 | 5.5712 |
| 4.2042 | 1.0 | 55 | 5.4376 |
| 3.7625 | 1.09 | 60 | 5.4613 |
| 3.5759 | 1.18 | 65 | 5.5305 |
| 3.6831 | 1.27 | 70 | 5.5329 |
| 3.7596 | 1.36 | 75 | 5.5254 |
| 3.6216 | 1.45 | 80 | 5.5825 |
| 3.769 | 1.55 | 85 | 5.6090 |
| 3.5107 | 1.64 | 90 | 5.6351 |
| 3.3485 | 1.73 | 95 | 5.6501 |
| 3.4216 | 1.82 | 100 | 5.6611 |
| 3.3527 | 1.91 | 105 | 5.7240 |
| 3.2204 | 2.0 | 110 | 5.8332 |
| 2.9853 | 2.09 | 115 | 5.8772 |
| 3.207 | 2.18 | 120 | 5.8846 |
| 3.4566 | 2.27 | 125 | 5.8788 |
| 3.1248 | 2.36 | 130 | 5.8898 |
| 3.0917 | 2.45 | 135 | 5.9108 |
| 3.1331 | 2.55 | 140 | 5.9545 |
| 2.9234 | 2.64 | 145 | 5.9664 |
| 3.0005 | 2.73 | 150 | 5.9582 |
| 3.4196 | 2.82 | 155 | 5.9526 |
| 3.2783 | 2.91 | 160 | 5.9486 |
| 3.1719 | 3.0 | 165 | 5.9468 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.1