metadata
library_name: transformers
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
metrics:
- wer
- bleu
model-index:
- name: whisper-small-OpenSLR-GL-EN
results: []
whisper-small-OpenSLR-GL-EN
This model is a fine-tuned version of openai/whisper-small on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.3434
- Wer: 30.6658
- Bleu: 70.1070
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: 1.25e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- total_train_batch_size: 32
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Bleu |
|---|---|---|---|---|---|
| 0.6437 | 1.0 | 131 | 0.5877 | 38.9863 | 55.0646 |
| 0.1773 | 2.0 | 262 | 0.4173 | 35.4619 | 65.1019 |
| 0.5521 | 3.0 | 393 | 0.3602 | 27.1687 | 69.0765 |
| 0.1255 | 4.0 | 524 | 0.3463 | 25.5246 | 70.7943 |
| 0.1548 | 5.0 | 655 | 0.3434 | 30.6658 | 70.1070 |
Framework versions
- Transformers 4.45.1
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0