learning_rate: 0.0001
train_batch_size: 8
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: 1000
num_epochs: 30
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
5.2657	0.7042	500	3.4052	1.0
3.4604	1.4085	1000	3.3970	1.0
3.4889	2.1127	1500	3.4595	1.0
3.5963	2.8169	2000	3.4579	1.0
3.607	3.5211	2500	3.4567	1.0
3.5932	4.2254	3000	3.5102	1.0
3.6017	4.9296	3500	3.5089	1.0
3.6086	5.6338	4000	3.4746	1.0
3.6014	6.3380	4500	3.4683	1.0
3.6007	7.0423	5000	3.4568	1.0
3.6099	7.7465	5500	3.4615	1.0
3.5995	8.4507	6000	3.4544	1.0
3.6109	9.1549	6500	3.5015	1.0
3.5982	9.8592	7000	3.4718	1.0
3.6124	10.5634	7500	3.4791	1.0
3.6113	11.2676	8000	3.4565	1.0
3.5949	11.9718	8500	3.4624	1.0
3.5874	12.6761	9000	3.4634	1.0
3.6002	13.3803	9500	3.4792	1.0
3.613	14.0845	10000	3.4799	1.0
3.6081	14.7887	10500	3.4619	1.0
3.6015	15.4930	11000	3.4570	1.0
3.5852	16.1972	11500	3.4621	1.0
3.6011	16.9014	12000	3.4601	1.0
3.59	17.6056	12500	3.4526	1.0
3.6148	18.3099	13000	3.4583	1.0
3.6036	19.0141	13500	3.4619	1.0
3.5981	19.7183	14000	3.4699	1.0
3.6325	20.4225	14500	3.4682	1.0
3.6172	21.1268	15000	3.4751	1.0
3.6495	21.8310	15500	3.4891	1.0
3.637	22.5352	16000	3.4879	1.0
3.6621	23.2394	16500	3.4929	1.0
3.6721	23.9437	17000	3.5014	1.0
3.6859	24.6479	17500	3.5191	1.0
3.6924	25.3521	18000	3.5274	1.0
3.6662	26.0563	18500	3.5423	1.0
3.6953	26.7606	19000	3.5657	1.0
3.6889	27.4648	19500	3.5983	1.0
3.7113	28.1690	20000	3.6738	1.0
3.7305	28.8732	20500	3.7047	1.0
3.763	29.5775	21000	3.7455	1.0

Hung1294/ASR-model-20k-01

ASR-model-20k-01