<!-- 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-onomatopoeia-finetune_smalldata_ESC50pretrained_3
This model is a fine-tuned version of /root/workspace/wav2vec2-pretrained_with_ESC50_10000epochs_32batch_2022-07-09_22-16-46/pytorch_model.bin on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.5350
- Cer: 1.2730
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: 64
- eval_batch_size: 16
- 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 | Cer |
|---|---|---|---|---|
| 4.4243 | 4.67 | 500 | 2.6901 | 1.1259 |
| 2.4282 | 9.35 | 1000 | 2.7495 | 1.1563 |
| 2.3377 | 14.02 | 1500 | 2.2475 | 0.9617 |
| 2.2434 | 18.69 | 2000 | 2.2765 | 1.1908 |
| 2.2731 | 23.36 | 2500 | 2.2574 | 1.1669 |
| 2.3436 | 28.04 | 3000 | 2.5350 | 1.2730 |
Framework versions
- Transformers 4.11.3
- Pytorch 1.9.1+cu111
- Datasets 1.13.3
- Tokenizers 0.10.3
