metadata
license: apache-2.0
base_model: openai/whisper-large-v2
tags:
- generated_from_trainer
model-index:
- name: finetune_base
results: []
finetune_base
This model is a fine-tuned version of openai/whisper-large-v2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2932
- Cer: 8.8686
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: 5e-06
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 0.0863 | 2.76 | 1000 | 0.2303 | 9.1216 |
| 0.0154 | 5.52 | 2000 | 0.2505 | 8.6437 |
| 0.002 | 8.29 | 3000 | 0.2877 | 8.6297 |
| 0.0021 | 11.05 | 4000 | 0.2932 | 8.8686 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.0.0
- Datasets 2.18.0
- Tokenizers 0.15.2