learning_rate: 0.0003
train_batch_size: 16
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
optimizer: Use 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: 400
num_epochs: 45
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
5.8553	1.0900	400	inf	0.8323	0.2422
0.7172	2.1801	800	inf	0.5075	0.1290
0.5066	3.2701	1200	inf	0.4640	0.1159
0.4291	4.3602	1600	inf	0.4259	0.1068
0.3576	5.4502	2000	inf	0.4220	0.1082
0.3094	6.5402	2400	inf	0.4083	0.1024
0.2778	7.6303	2800	inf	0.4011	0.0999
0.2332	8.7203	3200	inf	0.3973	0.0993
0.2108	9.8104	3600	inf	0.3815	0.0957
0.1844	10.9004	4000	inf	0.3934	0.0979
0.1651	11.9905	4400	inf	0.3929	0.0977
0.146	13.0791	4800	inf	0.3736	0.0946
0.1243	14.1692	5200	inf	0.3723	0.0941
0.1107	15.2592	5600	inf	0.3736	0.0943
0.1006	16.3492	6000	inf	0.3698	0.0936
0.0931	17.4393	6400	inf	0.3721	0.0939
0.0847	18.5293	6800	inf	0.3616	0.0924
0.0786	19.6194	7200	inf	0.3670	0.0923
0.0701	20.7094	7600	inf	0.3676	0.0924
0.0665	21.7995	8000	inf	0.3679	0.0920
0.064	22.8895	8400	inf	0.3599	0.0913
0.0574	23.9795	8800	inf	0.3596	0.0909
0.0544	25.0682	9200	inf	0.3556	0.0900
0.0529	26.1583	9600	inf	0.3603	0.0915
0.0501	27.2483	10000	inf	0.3534	0.0905
0.0474	28.3383	10400	inf	0.3508	0.0906
0.0421	29.4284	10800	inf	0.3619	0.0915
0.0411	30.5184	11200	inf	0.3561	0.0900
0.041	31.6085	11600	inf	0.3562	0.0909
0.0389	32.6985	12000	inf	0.3537	0.0899
0.0362	33.7885	12400	inf	0.3489	0.0892
0.0347	34.8786	12800	inf	0.3495	0.0884
0.0319	35.9686	13200	inf	0.3481	0.0891
0.031	37.0573	13600	inf	0.3426	0.0883
0.0292	38.1473	14000	inf	0.3428	0.0877
0.0288	39.2374	14400	inf	0.3409	0.0874
0.0273	40.3274	14800	inf	0.3421	0.0872
0.026	41.4175	15200	inf	0.3390	0.0866
0.0243	42.5075	15600	inf	0.3425	0.0876
0.0237	43.5975	16000	inf	0.3429	0.0877
0.0225	44.6876	16400	inf	0.3404	0.0875

kaarthu2003/wav2vec2-final-learning-rate-3e-4_batch-size-16-45-epochs

wav2vec2-final-learning-rate-3e-4_batch-size-16-45-epochs