<!-- 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. -->
Speech3
This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.9613
- Wer: 1
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: 0.01
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 3000
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
2.9722 | 0.19 | 100 | 3.0022 | 1 |
2.9665 | 0.38 | 200 | 2.9838 | 1 |
2.9826 | 0.56 | 300 | 2.9780 | 1 |
2.975 | 0.75 | 400 | 3.0293 | 1 |
3.0458 | 0.94 | 500 | 3.0058 | 1 |
2.9819 | 1.13 | 600 | 3.0088 | 1 |
3.0267 | 1.31 | 700 | 3.0086 | 1 |
2.945 | 1.5 | 800 | 2.9810 | 1 |
2.9501 | 1.69 | 900 | 2.9702 | 1 |
3.0021 | 1.88 | 1000 | 2.9810 | 1 |
2.9721 | 2.06 | 1100 | 3.0104 | 1 |
2.9669 | 2.25 | 1200 | 3.0150 | 1 |
2.9694 | 2.44 | 1300 | 2.9685 | 1 |
2.968 | 2.63 | 1400 | 3.0010 | 1 |
2.9604 | 2.81 | 1500 | 2.9753 | 1 |
3.0178 | 3.0 | 1600 | 3.2719 | 1 |
2.9699 | 3.19 | 1700 | 2.9617 | 1 |
2.9871 | 3.38 | 1800 | 2.9613 | 1 |
2.9551 | 3.56 | 1900 | 2.9946 | 1 |
2.9597 | 3.75 | 2000 | 2.9652 | 1 |
2.9573 | 3.94 | 2100 | 3.1862 | 1 |
2.9652 | 4.13 | 2200 | 3.0658 | 1 |
2.9627 | 4.32 | 2300 | 3.2517 | 1 |
2.9382 | 4.5 | 2400 | 2.9681 | 1 |
2.9273 | 4.69 | 2500 | 2.9715 | 1 |
2.9459 | 4.88 | 2600 | 3.0109 | 1 |
3.0296 | 5.07 | 2700 | 2.9807 | 1 |
3.0574 | 5.25 | 2800 | 2.9737 | 1 |
3.0118 | 5.44 | 2900 | 2.9675 | 1 |
2.9799 | 5.63 | 3000 | 2.9804 | 1 |
2.9839 | 5.82 | 3100 | 3.0447 | 1 |
2.9565 | 6.0 | 3200 | 2.9867 | 1 |
2.9251 | 6.19 | 3300 | 2.9778 | 1 |
2.9881 | 6.38 | 3400 | 3.2389 | 1 |
2.9388 | 6.57 | 3500 | 2.9655 | 1 |
2.9832 | 6.75 | 3600 | 3.0652 | 1 |
2.9741 | 6.94 | 3700 | 2.9748 | 1 |
2.9425 | 7.13 | 3800 | 2.9657 | 1 |
2.9116 | 7.32 | 3900 | 2.9609 | 1 |
2.9413 | 7.5 | 4000 | 2.9629 | 1 |
2.8981 | 7.69 | 4100 | 2.9748 | 1 |
2.9369 | 7.88 | 4200 | 2.9799 | 1 |
2.9941 | 8.07 | 4300 | 2.9553 | 1 |
2.97 | 8.26 | 4400 | 2.9607 | 1 |
2.9205 | 8.44 | 4500 | 2.9552 | 1 |
2.9585 | 8.63 | 4600 | 2.9680 | 1 |
2.9763 | 8.82 | 4700 | 2.9646 | 1 |
2.9206 | 9.01 | 4800 | 2.9615 | 1 |
2.9236 | 9.19 | 4900 | 2.9655 | 1 |
2.8985 | 9.38 | 5000 | 2.9613 | 1 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3