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
0.5761	1.0	2	1.6114	27.3705	100.5
0.5761	2.0	4	1.6114	27.5641	100.5338
0.5761	3.0	6	1.6115	27.6322	100.2635
0.5761	4.0	8	1.6123	27.5361	100.75
0.5761	5.0	10	1.6123	27.5029	100.4257
0.5761	6.0	12	1.6126	27.6782	100.277
0.5761	7.0	14	1.6134	27.6104	100.8581
0.5761	8.0	16	1.6140	27.1867	106.027
0.5761	9.0	18	1.6146	27.3661	100.4257
0.5761	10.0	20	1.6161	27.4141	100.7703
0.5761	11.0	22	1.6168	27.4184	101.1149
0.5761	12.0	24	1.6185	27.5589	100.8446
0.5761	13.0	26	1.6194	27.3075	100.0946
0.5761	14.0	28	1.6212	27.5114	101.2973
0.5761	15.0	30	1.6220	27.3325	100.2635
0.5761	16.0	32	1.6240	27.1967	100.5338
0.5761	17.0	34	1.6255	27.2817	100.7297
0.5761	18.0	36	1.6283	27.3868	99.9189
0.5761	19.0	38	1.6307	27.3481	99.6351
0.5761	20.0	40	1.6335	27.3647	99.5676
0.5761	21.0	42	1.6390	27.0167	99.5946

Tsedeniya/finetuned_glosslm_kmr_translit

finetuned_glosslm_kmr_translit