metadata
library_name: transformers
license: apache-2.0
base_model: openai/whisper-medium
tags:
- generated_from_trainer
datasets:
- toigen
metrics:
- wer
model-index:
- name: whisper-medium-toigen-female-model
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: toigen
type: toigen
metrics:
- name: Wer
type: wer
value: 0.44457831325301206
whisper-medium-toigen-female-model
This model is a fine-tuned version of openai/whisper-medium on the toigen dataset. It achieves the following results on the evaluation set:
- Loss: 0.5737
- Wer: 0.4446
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: 1e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.3083 | 1.6024 | 200 | 0.6687 | 0.5309 |
| 0.3381 | 3.2008 | 400 | 0.5737 | 0.4446 |
| 0.2639 | 4.8032 | 600 | 0.6060 | 0.4297 |
| 0.0831 | 6.4016 | 800 | 0.6427 | 0.4032 |
| 0.0776 | 8.0 | 1000 | 0.6705 | 0.4209 |
Framework versions
- Transformers 4.47.1
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0