<!-- 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-finetuned-ks
This model is a fine-tuned version of facebook/wav2vec2-base on the superb dataset. It achieves the following results on the evaluation set:
- Loss: 0.0812
- Accuracy: 0.9821
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: 3e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
0.4499 | 1.0 | 798 | 0.2799 | 0.9719 |
0.2284 | 2.0 | 1596 | 0.1266 | 0.9773 |
0.1911 | 3.0 | 2394 | 0.0990 | 0.9793 |
0.1759 | 4.0 | 3192 | 0.0867 | 0.9807 |
0.1119 | 5.0 | 3990 | 0.0812 | 0.9821 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1+cu116
- Datasets 1.14.0
- Tokenizers 0.13.2