exp28-connected-stage1-frozen

This model is a fine-tuned version of cyttic/trocr-hebrew-untrained on an unknown dataset. It achieves the following results on the evaluation set:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 16
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: 0.05
num_epochs: 1

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
5.4750	0.0499	2500	5.2921	0.7780	1.0498
5.1373	0.0997	5000	5.0355	0.8212	1.1426
4.9213	0.1496	7500	4.8952	0.8017	1.1251
4.8041	0.1995	10000	4.7412	0.7949	1.0785
4.6828	0.2493	12500	4.5593	0.7600	1.0510
4.5887	0.2992	15000	4.4055	0.7627	1.0710
4.3930	0.3491	17500	4.2588	0.7550	1.0665
4.1659	0.3989	20000	4.1075	0.7415	1.0555
4.1116	0.4488	22500	3.9941	0.7189	1.0115
3.9639	0.4987	25000	3.8622	0.7154	1.0110
3.9089	0.5485	27500	3.7470	0.7088	1.0091
3.7613	0.5984	30000	3.6414	0.6989	1.0056
3.7173	0.6483	32500	3.5645	0.6847	0.9785
3.5710	0.6981	35000	3.4901	0.6780	0.9697
3.5911	0.7480	37500	3.4246	0.6696	0.9554
3.4808	0.7979	40000	3.3439	0.6725	0.9748
3.4458	0.8477	42500	3.2941	0.6674	0.9600
3.3879	0.8976	45000	3.2523	0.6622	0.9569
3.3574	0.9475	47500	3.2160	0.6560	0.9540
3.3047	0.9973	50000	3.1991	0.6590	0.9521
3.3368	1.0	50133	3.1992	0.6595	0.9530