learning_rate: 5e-05
train_batch_size: 4
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 32
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: 100

Training Loss	Epoch	Step	Validation Loss	Chrf++	Gen Len
60.4727	1.0	2	2.6099	7.0432	188.8514
60.4727	2.0	4	2.6098	6.929	188.8649
60.4727	3.0	6	2.6070	6.9838	188.7703
60.4727	4.0	8	2.6096	6.9828	188.7973
60.4727	5.0	10	2.6089	6.9032	188.9054
60.4727	6.0	12	2.6078	6.9169	188.5811
60.4727	7.0	14	2.6082	7.0118	188.8378
60.4727	8.0	16	2.6078	6.9782	188.7703
60.4727	9.0	18	2.6077	7.0593	188.6959
60.4727	10.0	20	2.6057	7.055	188.8311
60.4727	11.0	22	2.6048	7.0099	188.7568
60.4727	12.0	24	2.6034	6.9204	188.527
60.4727	13.0	26	2.6006	6.8578	188.3716
60.4727	14.0	28	2.6013	6.8726	188.1824
60.4727	15.0	30	2.6011	6.8654	188.2162
60.4727	16.0	32	2.5987	6.7165	188.1757
60.4727	17.0	34	2.5976	6.6228	188.6216
60.4727	18.0	36	2.5967	6.7021	188.1959
60.4727	19.0	38	2.5929	6.6589	188.2703
60.4727	20.0	40	2.5900	6.2654	188.1149
60.4727	21.0	42	2.5857	6.3809	188.1689
60.4727	22.0	44	2.5828	6.3858	188.1554
60.4727	23.0	46	2.5784	6.388	187.7365
60.4727	24.0	48	2.5785	6.464	187.9595

Tsedeniya/finetuned_byT5large_kmr_translit

finetuned_byT5large_kmr_translit