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: 34
num_epochs: 3.0
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
0.9976	0.2957	34	1.0304	48.5974
0.5696	0.5913	68	0.5931	23.2099
0.5063	0.8870	102	0.5395	23.2192
0.4626	1.1826	136	0.5209	21.7253
0.4722	1.4783	170	0.5084	21.3506
0.4813	1.7739	204	0.5011	21.0650
0.4164	2.0696	238	0.4936	20.7840
0.432	2.3652	272	0.4927	20.7559
0.4153	2.6609	306	0.4903	20.6950
0.4359	2.9565	340	0.4889	20.7418

greenw0lf/whisper-new-nnat-20h-mmr-xlsr

whisper-new-nnat-20h-mmr-xlsr