Whisper Large v3 Turbo

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the MiZonal v2.0.0 dataset. It achieves the following results on the evaluation set:

Loss: 0.2491
Wer: 14.8314

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: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
training_steps: 6000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.409	0.66	500	0.3669	67.9012
0.2324	1.31	1000	0.3014	37.3105
0.1984	1.97	1500	0.2333	36.4239
0.133	2.62	2000	0.2324	37.5093
0.0763	3.28	2500	0.2268	23.2248
0.0784	3.93	3000	0.2237	18.8168
0.0465	4.59	3500	0.2238	16.2151
0.0258	5.24	4000	0.2370	17.4828
0.0233	5.9	4500	0.2323	15.3037
0.01	6.55	5000	0.2444	15.0302
0.0039	7.21	5500	0.2458	14.8065
0.0028	7.86	6000	0.2491	14.8314