learning_rate: 1e-05
train_batch_size: 16
eval_batch_size: 16
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: 10
training_steps: 3000
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.8621	100	0.6753	25.7669
No log	1.7241	200	0.6783	27.6074
0.5043	2.5862	300	0.7213	33.1288
0.5043	3.4483	400	0.7248	28.8344
0.5043	4.3103	500	0.7586	28.8344
0.0579	5.1724	600	0.7449	25.7669
0.0579	6.0345	700	0.9104	27.6074
0.0579	6.8966	800	0.8260	25.1534
0.0092	7.7586	900	0.8445	28.2209
0.0092	8.6207	1000	0.8575	28.2209
0.0092	9.4828	1100	0.8613	29.4479
0.0013	10.3448	1200	0.8621	25.1534
0.0013	11.2069	1300	0.8488	26.3804
0.0013	12.0690	1400	0.8526	26.3804
0.0001	12.9310	1500	0.8537	26.3804
0.0001	13.7931	1600	0.8559	26.9939
0.0001	14.6552	1700	0.8573	26.3804
0.0001	15.5172	1800	0.8594	26.3804
0.0001	16.3793	1900	0.8611	26.9939
0.0001	17.2414	2000	0.8616	26.9939
0.0001	18.1034	2100	0.8627	26.9939
0.0001	18.9655	2200	0.8634	26.3804
0.0001	19.8276	2300	0.8646	26.9939
0.0	20.6897	2400	0.8647	26.9939
0.0	21.5517	2500	0.8653	26.9939
0.0	22.4138	2600	0.8658	26.9939
0.0	23.2759	2700	0.8661	26.9939
0.0	24.1379	2800	0.8663	26.9939
0.0	25.0	2900	0.8664	26.9939
0.0	25.8621	3000	0.8665	26.9939

KaleeswaranM/model_v3

WhisperSmall