<!-- 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. -->
wav2vec2-xlsr-ft-cy-btb
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the TECHIAITH/BANC-TRAWSGRIFIADAU-BANGOR - CY dataset. It achieves the following results on the evaluation set:
- Loss: 2.3938
- Wer: 0.9983
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: 0.0003
- train_batch_size: 32
- eval_batch_size: 128
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 800
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
No log | 0.33 | 100 | 3.7176 | 1.0 |
No log | 0.66 | 200 | 3.0649 | 1.0 |
No log | 1.0 | 300 | 2.9714 | 1.0 |
No log | 1.33 | 400 | 1.3840 | 0.8891 |
3.628 | 1.66 | 500 | 1.0746 | 0.7500 |
3.628 | 1.99 | 600 | 0.8992 | 0.6832 |
3.628 | 2.33 | 700 | 0.8292 | 0.6175 |
3.628 | 2.66 | 800 | 0.8293 | 0.6080 |
3.628 | 2.99 | 900 | 0.6632 | 0.5244 |
0.8635 | 3.32 | 1000 | 0.6550 | 0.4947 |
0.8635 | 3.65 | 1100 | 0.6255 | 0.4777 |
0.8635 | 3.99 | 1200 | 0.5891 | 0.4465 |
0.8635 | 4.32 | 1300 | 0.7092 | 0.5183 |
0.8635 | 4.65 | 1400 | 1.0072 | 0.6188 |
0.8316 | 4.98 | 1500 | 1.2552 | 0.9375 |
0.8316 | 5.32 | 1600 | 1.4598 | 0.8563 |
0.8316 | 5.65 | 1700 | 1.6682 | 0.8975 |
0.8316 | 5.98 | 1800 | 1.8454 | 0.9660 |
0.8316 | 6.31 | 1900 | 2.0722 | 0.9919 |
1.8552 | 6.64 | 2000 | 2.0574 | 0.9931 |
1.8552 | 6.98 | 2100 | 2.1060 | 0.9887 |
1.8552 | 7.31 | 2200 | 2.0413 | 0.9771 |
1.8552 | 7.64 | 2300 | 1.9971 | 0.9765 |
1.8552 | 7.97 | 2400 | 1.9847 | 0.9742 |
2.1436 | 8.31 | 2500 | 2.0486 | 0.9700 |
2.1436 | 8.64 | 2600 | 2.1406 | 0.9661 |
2.1436 | 8.97 | 2700 | 2.3666 | 0.9527 |
2.1436 | 9.3 | 2800 | 2.2622 | 0.9692 |
2.1436 | 9.63 | 2900 | 2.2790 | 0.9943 |
2.3335 | 9.97 | 3000 | 2.3607 | 0.9972 |
2.3335 | 10.3 | 3100 | 2.3924 | 0.9983 |
2.3335 | 10.63 | 3200 | 2.3938 | 0.9983 |
2.3335 | 10.96 | 3300 | 2.3938 | 0.9983 |
2.3335 | 11.3 | 3400 | 2.3938 | 0.9983 |
2.4814 | 11.63 | 3500 | 2.3938 | 0.9983 |
2.4814 | 11.96 | 3600 | 2.3938 | 0.9983 |
2.4814 | 12.29 | 3700 | 2.3938 | 0.9983 |
2.4814 | 12.62 | 3800 | 2.3938 | 0.9983 |
2.4814 | 12.96 | 3900 | 2.3938 | 0.9983 |
2.4882 | 13.29 | 4000 | 2.3938 | 0.9983 |
2.4882 | 13.62 | 4100 | 2.3938 | 0.9983 |
2.4882 | 13.95 | 4200 | 2.3938 | 0.9983 |
2.4882 | 14.29 | 4300 | 2.3938 | 0.9983 |
2.4882 | 14.62 | 4400 | 2.3938 | 0.9983 |
2.4818 | 14.95 | 4500 | 2.3938 | 0.9983 |
2.4818 | 15.28 | 4600 | 2.3938 | 0.9983 |
2.4818 | 15.61 | 4700 | 2.3938 | 0.9983 |
2.4818 | 15.95 | 4800 | 2.3938 | 0.9983 |
2.4818 | 16.28 | 4900 | 2.3938 | 0.9983 |
2.487 | 16.61 | 5000 | 2.3938 | 0.9983 |
Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3