learning_rate: 0.0001
train_batch_size: 48
eval_batch_size: 32
seed: 42
optimizer: Use OptimizerNames.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: 11
num_epochs: 3.0
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
1.0515	0.3056	11	1.1412	36.7229
0.8641	0.6111	22	0.8851	33.7001
0.6831	0.9167	33	0.6276	26.5776
0.5236	1.2222	44	0.5228	22.1525
0.4989	1.5278	55	0.4746	21.7768
0.4472	1.8333	66	0.4490	21.0387
0.517	2.1389	77	0.4359	20.6730
0.5273	2.4444	88	0.4292	20.3778
0.39	2.75	99	0.4256	20.7032

greenw0lf/whisper-5hrs-random-seed333

whisper-5hrs-random-seed333