learning_rate: 5e-05
train_batch_size: 32
eval_batch_size: 16
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: linear
lr_scheduler_warmup_steps: 1500
training_steps: 15000

Training Loss	Epoch	Step	Validation Loss	Wer	Wer Ortho	Cer
2.9948	0.0344	100	1.6092	96.8735	97.4346	32.7676
1.1299	0.0689	200	0.9816	77.2371	78.9324	23.9400
0.8042	0.1033	300	0.8131	68.7561	70.9141	21.3040
0.6818	0.1378	400	0.7165	62.1701	64.7500	19.6127
0.6024	0.1722	500	0.6621	57.3447	59.9550	17.5739
0.5637	0.2067	600	0.6193	54.6983	57.9702	17.3628
0.5122	0.2411	700	0.5929	51.5446	54.5785	16.0688
0.4843	0.2756	800	0.5676	50.2459	53.2958	16.2271
0.4753	0.3100	900	0.5539	47.8818	51.1292	15.2683
0.4569	0.3445	1000	0.5373	47.7920	50.7530	15.8434
0.4531	0.3789	1100	0.5112	45.7182	48.8753	14.9207
0.4318	0.4134	1200	0.5019	42.7152	45.8653	13.5349
0.4189	0.4478	1300	0.4891	43.5510	46.5295	14.0562
0.4220	0.4823	1400	0.4889	42.8368	45.7590	13.8072
0.4069	0.5167	1500	0.4851	42.1652	45.5889	13.6825
0.3934	0.5512	1600	0.4796	41.7677	44.7581	13.9297
0.3791	0.5856	1700	0.4602	39.5379	42.5186	12.4823
0.3737	0.6200	1800	0.4502	39.6395	42.7949	12.8918
0.3783	0.6545	1900	0.4438	38.5559	41.4105	12.8791
0.3573	0.6889	2000	0.4364	38.8881	41.9353	13.1695
0.3521	0.7234	2100	0.4225	36.7508	39.7552	11.9780
0.3413	0.7578	2200	0.4220	36.9378	39.9676	11.9971
0.3442	0.7923	2300	0.4144	36.9986	39.9541	12.0782
0.3290	0.8267	2400	0.4102	36.0783	39.0702	11.6018
0.3219	0.8612	2500	0.4040	35.6291	38.5571	11.9874
0.3262	0.8956	2600	0.3961	34.8559	37.7416	11.2353
0.3168	0.9301	2700	0.3946	34.3821	37.3950	11.4266
0.3150	0.9645	2800	0.3851	34.4076	37.2591	11.1372
0.3099	0.9990	2900	0.3907	34.8713	37.7560	11.2911
0.2505	1.0334	3000	0.3853	34.8241	37.7641	11.7109

ihanif/exp_008b_base_cv24_vanilla_no_aug_low_wer

exp_008b_base_cv24_vanilla_no_aug