<!-- 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-base-timit-small
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: 0.5361
- Wer: 0.3380
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: 8
- 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
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
3.571 | 1.0 | 500 | 1.9252 | 1.0022 |
0.8969 | 2.01 | 1000 | 0.5066 | 0.5292 |
0.4326 | 3.01 | 1500 | 0.4523 | 0.4562 |
0.2993 | 4.02 | 2000 | 0.4228 | 0.4202 |
0.2335 | 5.02 | 2500 | 0.4252 | 0.4178 |
0.2009 | 6.02 | 3000 | 0.4136 | 0.3910 |
0.1552 | 7.03 | 3500 | 0.4747 | 0.3863 |
0.1388 | 8.03 | 4000 | 0.4359 | 0.3859 |
0.1226 | 9.04 | 4500 | 0.4367 | 0.3879 |
0.1109 | 10.04 | 5000 | 0.4360 | 0.3760 |
0.0991 | 11.04 | 5500 | 0.4899 | 0.3672 |
0.0882 | 12.05 | 6000 | 0.4608 | 0.3653 |
0.0792 | 13.05 | 6500 | 0.4882 | 0.3703 |
0.0745 | 14.06 | 7000 | 0.4716 | 0.3625 |
0.065 | 15.06 | 7500 | 0.4896 | 0.3651 |
0.0596 | 16.06 | 8000 | 0.4831 | 0.3659 |
0.0563 | 17.07 | 8500 | 0.5092 | 0.3585 |
0.0536 | 18.07 | 9000 | 0.5376 | 0.3675 |
0.0465 | 19.08 | 9500 | 0.5019 | 0.3534 |
0.049 | 20.08 | 10000 | 0.4869 | 0.3723 |
0.0423 | 21.08 | 10500 | 0.4947 | 0.3501 |
0.0348 | 22.09 | 11000 | 0.5524 | 0.3453 |
0.0315 | 23.09 | 11500 | 0.5369 | 0.3499 |
0.0312 | 24.1 | 12000 | 0.5283 | 0.3519 |
0.0258 | 25.1 | 12500 | 0.5202 | 0.3461 |
0.0249 | 26.1 | 13000 | 0.5270 | 0.3449 |
0.0236 | 27.11 | 13500 | 0.5388 | 0.3408 |
0.0206 | 28.11 | 14000 | 0.5361 | 0.3388 |
0.0224 | 29.12 | 14500 | 0.5361 | 0.3380 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.13.1+cu116
- Datasets 1.18.3
- Tokenizers 0.13.2