<!-- 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. -->
retrain_epoch2to5
This model is a fine-tuned version of alexziweiwang/retrain_first1epoch on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 3.3244
- Acc: 0.24
- Wer: 1.0
- Correct: 48
- Total: 200
- Strlen: 200
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: 9e-06
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1.0
Training results
Training Loss | Epoch | Step | Validation Loss | Acc | Wer | Correct | Total | Strlen |
---|---|---|---|---|---|---|---|---|
No log | 0.02 | 5 | 7.8494 | 0.24 | 1.0 | 48 | 200 | 200 |
7.6032 | 0.04 | 10 | 7.4834 | 0.24 | 1.0 | 48 | 200 | 200 |
7.6032 | 0.06 | 15 | 7.1350 | 0.24 | 1.0 | 48 | 200 | 200 |
7.3336 | 0.08 | 20 | 6.8284 | 0.24 | 1.0 | 48 | 200 | 200 |
7.3336 | 0.11 | 25 | 6.5577 | 0.24 | 1.0 | 48 | 200 | 200 |
6.2911 | 0.13 | 30 | 6.3124 | 0.24 | 1.0 | 48 | 200 | 200 |
6.2911 | 0.15 | 35 | 6.0850 | 0.24 | 1.0 | 48 | 200 | 200 |
5.9181 | 0.17 | 40 | 5.8888 | 0.24 | 1.0 | 48 | 200 | 200 |
5.9181 | 0.19 | 45 | 5.6815 | 0.24 | 1.0 | 48 | 200 | 200 |
5.7954 | 0.21 | 50 | 5.4834 | 0.24 | 1.0 | 48 | 200 | 200 |
5.7954 | 0.23 | 55 | 5.3099 | 0.24 | 1.0 | 48 | 200 | 200 |
5.4801 | 0.25 | 60 | 5.1678 | 0.24 | 1.0 | 48 | 200 | 200 |
5.4801 | 0.27 | 65 | 5.0223 | 0.24 | 1.0 | 48 | 200 | 200 |
5.3377 | 0.3 | 70 | 4.8893 | 0.24 | 1.0 | 48 | 200 | 200 |
5.3377 | 0.32 | 75 | 4.7743 | 0.24 | 1.0 | 48 | 200 | 200 |
5.2511 | 0.34 | 80 | 4.6494 | 0.24 | 1.0 | 48 | 200 | 200 |
5.2511 | 0.36 | 85 | 4.5307 | 0.24 | 1.0 | 48 | 200 | 200 |
4.727 | 0.38 | 90 | 4.4237 | 0.24 | 1.0 | 48 | 200 | 200 |
4.727 | 0.4 | 95 | 4.3263 | 0.24 | 1.0 | 48 | 200 | 200 |
4.7653 | 0.42 | 100 | 4.2439 | 0.24 | 1.0 | 48 | 200 | 200 |
4.7653 | 0.44 | 105 | 4.1589 | 0.24 | 1.0 | 48 | 200 | 200 |
4.4971 | 0.46 | 110 | 4.0847 | 0.24 | 1.0 | 48 | 200 | 200 |
4.4971 | 0.48 | 115 | 4.0118 | 0.24 | 1.0 | 48 | 200 | 200 |
4.0077 | 0.51 | 120 | 3.9382 | 0.24 | 1.0 | 48 | 200 | 200 |
4.0077 | 0.53 | 125 | 3.8663 | 0.24 | 1.0 | 48 | 200 | 200 |
4.1693 | 0.55 | 130 | 3.8106 | 0.24 | 1.0 | 48 | 200 | 200 |
4.1693 | 0.57 | 135 | 3.7580 | 0.24 | 1.0 | 48 | 200 | 200 |
4.0854 | 0.59 | 140 | 3.7123 | 0.24 | 1.0 | 48 | 200 | 200 |
4.0854 | 0.61 | 145 | 3.6720 | 0.24 | 1.0 | 48 | 200 | 200 |
4.1988 | 0.63 | 150 | 3.6260 | 0.24 | 1.0 | 48 | 200 | 200 |
4.1988 | 0.65 | 155 | 3.5853 | 0.24 | 1.0 | 48 | 200 | 200 |
3.9975 | 0.67 | 160 | 3.5463 | 0.24 | 1.0 | 48 | 200 | 200 |
3.9975 | 0.7 | 165 | 3.5122 | 0.24 | 1.0 | 48 | 200 | 200 |
3.6042 | 0.72 | 170 | 3.4862 | 0.24 | 1.0 | 48 | 200 | 200 |
3.6042 | 0.74 | 175 | 3.4631 | 0.24 | 1.0 | 48 | 200 | 200 |
3.7347 | 0.76 | 180 | 3.4406 | 0.24 | 1.0 | 48 | 200 | 200 |
3.7347 | 0.78 | 185 | 3.4202 | 0.24 | 1.0 | 48 | 200 | 200 |
3.8336 | 0.8 | 190 | 3.4014 | 0.24 | 1.0 | 48 | 200 | 200 |
3.8336 | 0.82 | 195 | 3.3855 | 0.24 | 1.0 | 48 | 200 | 200 |
3.7454 | 0.84 | 200 | 3.3703 | 0.24 | 1.0 | 48 | 200 | 200 |
3.7454 | 0.86 | 205 | 3.3576 | 0.24 | 1.0 | 48 | 200 | 200 |
3.525 | 0.89 | 210 | 3.3471 | 0.24 | 1.0 | 48 | 200 | 200 |
3.525 | 0.91 | 215 | 3.3392 | 0.24 | 1.0 | 48 | 200 | 200 |
3.8175 | 0.93 | 220 | 3.3331 | 0.24 | 1.0 | 48 | 200 | 200 |
3.8175 | 0.95 | 225 | 3.3289 | 0.24 | 1.0 | 48 | 200 | 200 |
3.307 | 0.97 | 230 | 3.3259 | 0.24 | 1.0 | 48 | 200 | 200 |
3.307 | 0.99 | 235 | 3.3244 | 0.24 | 1.0 | 48 | 200 | 200 |
Framework versions
- Transformers 4.23.1
- Pytorch 1.12.1+cu113
- Datasets 1.18.3
- Tokenizers 0.13.2