learning_rate: 5e-05
train_batch_size: 6
eval_batch_size: 6
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
num_epochs: 30
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Cer
1.5513	0.8947	8333	1.3010	0.5588
0.4355	1.7893	16666	0.4287	0.2111
0.1775	2.6840	24999	0.2173	0.1193
0.0963	3.5787	33332	0.1354	0.0758
0.0613	4.4734	41665	0.1058	0.0505
0.0389	5.3680	49998	0.0990	0.0573
0.0257	6.2627	58331	0.0873	0.0439
0.0198	7.1574	66664	0.0822	0.0377
0.0256	8.0521	74997	0.0690	0.0319
0.0155	8.9467	83330	0.0692	0.0457
0.013	9.8414	91663	0.0583	0.0315
0.0145	10.7361	99996	0.0647	0.0304
0.0128	11.6308	108329	0.0601	0.0405
0.0098	12.5254	116662	0.0527	0.0285
0.0093	13.4201	124995	0.0585	0.0244
0.0117	14.3148	133328	0.0518	0.0334
0.0054	15.2095	141661	0.0502	0.0361
0.0064	16.1041	149994	0.0479	0.0245
0.0039	16.9988	158327	0.0410	0.0337
0.004	17.8935	166660	0.0498	0.0366
0.007	18.7882	174993	0.0488	0.0362
0.0027	19.6828	183326	0.0471	0.0272
0.001	20.5775	191659	0.0407	0.0270
0.0011	21.4722	199992	0.0450	0.0401
0.0016	22.3669	208325	0.0468	0.0343
0.0004	23.2615	216658	0.0393	0.0245
0.0004	24.1562	224991	0.0427	0.0227
0.0012	25.0509	233324	0.0380	0.0241
0.0003	25.9456	241657	0.0359	0.0224
0.0004	26.8402	249990	0.0370	0.0260
0.0008	27.7349	258323	0.0363	0.0321
0.0004	28.6296	266656	0.0361	0.0219
0.0	29.5243	274989	0.0364	0.0229

amanladia/results_16nov

results_16nov