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
11.6680	1.0	2	1.2881	28.25	71.375
11.6680	2.0	4	1.2851	28.7668	71.25
11.6680	3.0	6	1.2806	28.3142	70.9375
11.6680	4.0	8	1.2740	28.3785	70.875
11.6680	5.0	10	1.2677	28.4686	70.5625
11.6680	6.0	12	1.2577	28.4344	71.375
11.6680	7.0	14	1.2470	28.3743	69.9375
11.6680	8.0	16	1.2364	28.0997	70.6875
11.6680	9.0	18	1.2240	28.3145	69.125
11.6680	10.0	20	1.2100	28.103	68.5625
11.6680	11.0	22	1.1960	28.103	68.5625
11.6680	12.0	24	1.1806	27.2219	68.3125
11.6680	13.0	26	1.1656	27.4538	68.75
11.6680	14.0	28	1.1488	28.0965	68.0625
11.6680	15.0	30	1.1318	27.5217	67.1875
11.6680	16.0	32	1.1153	27.565	66.5
11.6680	17.0	34	1.0967	28.6044	67.6875
11.6680	18.0	36	1.0788	28.6629	69.25
11.6680	19.0	38	1.0614	28.1279	67.4375

Tsedeniya/finetuned_glosslm_pan_translit

finetuned_glosslm_pan_translit