learning_rate: 0.0003
train_batch_size: 4
eval_batch_size: 4
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: constant
num_epochs: 6
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.0711	0.1946	100	3.0752	1.0	1.0
2.9832	0.3891	200	2.9191	1.0	0.9808
2.3637	0.5837	300	2.0488	1.0	0.6432
1.5098	0.7782	400	1.2189	0.9996	0.3750
1.0659	0.9728	500	0.8497	0.8824	0.2449
0.8263	1.1673	600	0.6494	0.7365	0.1706
0.7036	1.3619	700	0.5537	0.6918	0.1511
0.7094	1.5564	800	0.5145	0.6266	0.1361
0.6911	1.7510	900	0.5001	0.6041	0.1286
0.6203	1.9455	1000	0.4466	0.5762	0.1217
0.5097	2.1401	1100	0.4221	0.5713	0.1122
0.5037	2.3346	1200	0.4455	0.5836	0.1228
0.474	2.5292	1300	0.4046	0.5283	0.1058
0.4551	2.7237	1400	0.3767	0.4992	0.0987
0.4929	2.9183	1500	0.3976	0.5	0.1000
0.393	3.1128	1600	0.3882	0.5201	0.1027
0.409	3.3074	1700	0.3830	0.4971	0.0959
0.4253	3.5019	1800	0.3815	0.5205	0.1007
0.4148	3.6965	1900	0.3754	0.5078	0.0968
0.3601	3.8911	2000	0.3541	0.4676	0.0898
0.3241	4.0856	2100	0.3672	0.4906	0.0929
0.3262	4.2802	2200	0.3598	0.4594	0.0884
0.3535	4.4747	2300	0.4117	0.5078	0.0976
0.3519	4.6693	2400	0.3565	0.4512	0.0874
0.3408	4.8638	2500	0.3440	0.4369	0.0863
0.3596	5.0584	2600	0.3581	0.4344	0.0844
0.3338	5.2529	2700	0.3963	0.4611	0.0875
0.4681	5.4475	2800	0.5356	0.5898	0.1142
0.5929	5.6420	2900	0.6560	0.6820	0.1454
0.8122	5.8366	3000	1.0430	0.8525	0.2006

tadgeis/xls-r-300m-ckt-staged-bible-radio-to-chuklang-source-eval-stage1

xls-r-300m-ckt-staged-bible-radio-to-chuklang-source-eval-stage1