metadata
license: apache-2.0
base_model: facebook/wav2vec2-base-960h
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: 960h
results: []
960h
This model is a fine-tuned version of facebook/wav2vec2-base-960h on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.4593
- Wer: 39.9917
- Cer: 21.1881
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: 32
- 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: 2000
- training_steps: 5000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 4.9803 | 18.52 | 1000 | 2.7949 | 100.0 | 98.8206 |
| 1.1136 | 37.04 | 2000 | 1.2272 | 41.3621 | 21.5532 |
| 0.4547 | 55.56 | 3000 | 1.3323 | 41.2514 | 21.6285 |
| 0.2591 | 74.07 | 4000 | 1.3635 | 39.3688 | 20.8403 |
| 0.2232 | 92.59 | 5000 | 1.4593 | 39.9917 | 21.1881 |
Framework versions
- Transformers 4.39.3
- Pytorch 1.12.1+cu116
- Datasets 2.18.0
- Tokenizers 0.15.2