learning_rate: 5e-05
train_batch_size: 32
eval_batch_size: 32
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 64
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
lr_scheduler_warmup_steps: 200
num_epochs: 15
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf
4.8353	1.0	115	1.7278	10.32	32.32
3.2422	2.0	230	1.1679	21.54	43.41
2.4046	3.0	345	0.9479	28.04	48.77
1.9010	4.0	460	0.8590	30.8	51.59
1.6190	5.0	575	0.8173	32.09	52.83
1.4331	6.0	690	0.7880	34.24	54.57
1.2747	7.0	805	0.7756	35.33	55.51
1.1539	8.0	920	0.7649	35.8	55.89
1.0629	9.0	1035	0.7607	35.98	56.13
0.9790	10.0	1150	0.7565	36.88	56.7
0.9260	11.0	1265	0.7579	36.81	56.79
0.8740	12.0	1380	0.7579	36.82	56.86
0.8311	13.0	1495	0.7584	37.55	57.25
0.8215	14.0	1610	0.7585	37.41	57.24
0.7909	15.0	1725	0.7597	37.52	57.39

eliezermga/ruund-fr-nllb

ruund-fr-nllb