learning_rate: 0.0001
train_batch_size: 48
eval_batch_size: 32
seed: 42
optimizer: Use adamw_torch_fused 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: 19
num_epochs: 3.0
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
1.0761	0.2308	15	1.2218	42.8038
0.7456	0.4615	30	0.8029	35.4863
0.5595	0.6923	45	0.6174	20.2165
0.4555	0.9231	60	0.5368	17.4351
0.4447	1.1538	75	0.5081	18.1445
0.4297	1.3846	90	0.4910	16.5111
0.4335	1.6154	105	0.4824	16.3151
0.4193	1.8462	120	0.4754	16.0258
0.3978	2.0769	135	0.4696	16.0818
0.4019	2.3077	150	0.4656	16.0724
0.4146	2.5385	165	0.4629	16.0071
0.4415	2.7692	180	0.4612	16.0724
0.3733	3.0	195	0.4604	16.0911

greenw0lf/whisper-nonnat-10h-random-seed42

whisper-nonnat-10h-random-seed42