This model is a fine-tuned version of openai/whisper-small on the custom_torgo_0_0 dataset merged with the UASpeech dataset.
It achieves the following results on the evaluation set:
Wer: 34.5269
Wer Ortho: 36.0478
And the following results on the TORGO + UAS training set:
Acc: 0.68
Wer: 32.28
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
learning_rate: 1e-05
train_batch_size: 8
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
