learning_rate: 1e-05
train_batch_size: 4
eval_batch_size: 8
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: 50
training_steps: 1700
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
0.5424	0.4854	100	0.6264	65.1653
0.4626	0.9709	200	0.5054	56.4463
0.2254	1.4563	300	0.4883	57.9752
0.1756	1.9417	400	0.4684	52.7273
0.0827	2.4272	500	0.4958	52.1901
0.0579	2.9126	600	0.4807	50.9091
0.0233	3.3981	700	0.5169	50.5372
0.0175	3.8835	800	0.5269	49.1736
0.0061	4.3689	900	0.5338	47.8099
0.007	4.8544	1000	0.5347	50.0
0.0021	5.3398	1100	0.5416	47.8926
0.0034	5.8252	1200	0.5490	49.2562
0.0014	6.3107	1300	0.5583	47.7686
0.0011	6.7961	1400	0.5583	47.0661
0.001	7.2816	1500	0.5605	46.9008
0.0008	7.7670	1600	0.5632	47.0248
0.0008	8.2524	1700	0.5644	47.0248

EdBergJr/whisper-small-hi

Whisper Small Hi - Sanchit Gandhi