<!-- 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. -->
torgo-TestedSpeakerM01-finetuned-On-torgoSentenceModel
This model is a fine-tuned version of alexziweiwang/torgo-sentences-TestedSpeakerM01 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3376
- Wer: 0.7172
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: 0.0001
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 30
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
15.6854 | 0.84 | 500 | 0.4620 | 1.0596 |
0.3374 | 1.69 | 1000 | 0.2646 | 1.0921 |
0.3121 | 2.53 | 1500 | 0.2602 | 1.0704 |
0.2683 | 3.37 | 2000 | 0.2417 | 1.0997 |
0.2478 | 4.22 | 2500 | 0.2438 | 1.0217 |
0.2368 | 5.06 | 3000 | 0.2600 | 0.9729 |
0.2236 | 5.9 | 3500 | 0.2400 | 0.9112 |
0.2183 | 6.75 | 4000 | 0.3123 | 0.9805 |
0.1989 | 7.59 | 4500 | 0.2716 | 0.9231 |
0.1852 | 8.43 | 5000 | 0.2539 | 0.8732 |
0.1778 | 9.27 | 5500 | 0.2673 | 0.8516 |
0.1769 | 10.12 | 6000 | 0.2951 | 0.8819 |
0.1798 | 10.96 | 6500 | 0.3194 | 0.9122 |
0.1525 | 11.8 | 7000 | 0.2320 | 0.8830 |
0.1497 | 12.65 | 7500 | 0.2814 | 0.8007 |
0.1401 | 13.49 | 8000 | 0.2693 | 0.8039 |
0.1227 | 14.33 | 8500 | 0.2681 | 0.7887 |
0.1259 | 15.18 | 9000 | 0.2592 | 0.7584 |
0.1278 | 16.02 | 9500 | 0.2795 | 0.7920 |
0.1124 | 16.86 | 10000 | 0.3190 | 0.7714 |
0.1116 | 17.71 | 10500 | 0.3306 | 0.7909 |
0.1024 | 18.55 | 11000 | 0.2935 | 0.7519 |
0.1417 | 19.39 | 11500 | 0.3441 | 0.7638 |
0.0924 | 20.24 | 12000 | 0.3136 | 0.7454 |
0.0916 | 21.08 | 12500 | 0.3144 | 0.7107 |
0.0875 | 21.92 | 13000 | 0.3350 | 0.7421 |
0.087 | 22.77 | 13500 | 0.3129 | 0.7530 |
0.0741 | 23.61 | 14000 | 0.3384 | 0.7454 |
0.0708 | 24.45 | 14500 | 0.3688 | 0.7486 |
0.0812 | 25.3 | 15000 | 0.3383 | 0.7216 |
0.0669 | 26.14 | 15500 | 0.3201 | 0.7205 |
0.0612 | 26.98 | 16000 | 0.3116 | 0.7140 |
0.0637 | 27.82 | 16500 | 0.3303 | 0.7064 |
0.0629 | 28.67 | 17000 | 0.3382 | 0.7129 |
0.0667 | 29.51 | 17500 | 0.3350 | 0.7161 |
Framework versions
- Transformers 4.23.1
- Pytorch 1.12.1+cu113
- Datasets 1.18.3
- Tokenizers 0.13.2