learning_rate: 3e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 64
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: 400
num_epochs: 25
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Wer
1.8047	0.9791	41	1.7156	0.0044	0.3448
1.1095	1.9791	82	1.1373	0.0044	0.2618
0.7728	2.9791	123	0.9137	0.0044	0.2129
0.617	3.9791	164	0.7351	0.0044	0.1931
0.3968	4.9791	205	0.5717	0.0044	0.1985
0.2369	5.9791	246	0.5087	0.0044	0.1889
0.1676	6.9791	287	0.5160	0.0044	0.1728
0.1168	7.9791	328	0.5312	0.0044	0.1717
0.0758	8.9791	369	0.5481	0.0044	0.1845
0.0468	9.9791	410	0.5740	0.0044	0.1895
0.035	10.9791	451	0.5997	0.0044	0.1994
0.0173	11.9791	492	0.6203	0.0044	0.1831
0.0156	12.9791	533	0.6272	0.0044	0.1976
0.0096	13.9791	574	0.6369	0.0044	0.1849
0.0064	14.9791	615	0.6480	0.0044	0.2024
0.004	15.9791	656	0.6613	0.0044	0.1870
0.0029	16.9791	697	0.6676	0.0044	0.1982
0.0023	17.9791	738	0.6830	0.0044	0.1865
0.0029	18.9791	779	0.6858	0.0044	0.1872
0.0018	19.9791	820	0.6931	0.0044	0.1999
0.0018	20.9791	861	0.6942	0.0044	0.1860
0.0018	21.9791	902	0.6998	0.0044	0.1854
0.0016	22.9791	943	0.6995	0.0044	0.1861
0.0016	23.9791	984	0.7011	0.0044	0.1858
0.0016	24.9791	1025	0.6991	0.0044	0.1857

mdsingh2024/ap-OWxIIz49ySv6E8XQhYeXAB

ap-OWxIIz49ySv6E8XQhYeXAB