learning_rate: 4e-05
train_batch_size: 16
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: constant_with_warmup
num_epochs: 7
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
No log	1.0	14	0.7929	0.3230	0.3224
0.0136	2.0	28	0.7854	0.3378	0.3372
0.0136	3.0	42	0.7646	0.2907	0.2868
0.0102	4.0	56	0.8354	0.2914	0.2901
0.0102	5.0	70	0.7656	0.3244	0.3230
0.0103	6.0	84	0.8051	0.2968	0.2933
0.0103	7.0	98	0.8514	0.3318	0.3301

bogosla/whisper-tiny-en

whisper-tiny-en