learning_rate: 2e-05
train_batch_size: 16
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
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
num_epochs: 5
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Bleu	Meteor	Ter
5.7226	0.16	500	1.2374	33.8521	0.5919	52.9543
4.6943	0.32	1000	1.0143	37.6161	0.6215	50.2010
4.1820	0.48	1500	0.9243	38.7348	0.6297	48.9987
3.9617	0.64	2000	0.8697	39.6214	0.6372	48.2792
3.8142	0.8	2500	0.8348	39.9645	0.6402	48.1631
3.6022	0.96	3000	0.8088	40.4316	0.6424	47.5672
3.4027	1.12	3500	0.7904	40.8243	0.6464	47.4910
3.2837	1.28	4000	0.7767	40.9658	0.6466	47.3393
3.3361	1.44	4500	0.7636	41.0778	0.6473	47.1045
3.2255	1.6	5000	0.7519	41.3600	0.6492	46.9941
3.1059	1.76	5500	0.7440	41.4582	0.6506	46.8724
3.1077	1.92	6000	0.7354	41.5073	0.6496	46.8599
2.9352	2.08	6500	0.7306	41.5449	0.6505	46.8121
2.9681	2.24	7000	0.7249	41.7274	0.6514	46.7468
2.9900	2.4	7500	0.7206	41.7396	0.6516	46.6186
2.8702	2.56	8000	0.7163	41.9143	0.6523	46.5561
2.8863	2.7200	8500	0.7137	41.9456	0.6536	46.4100
2.8744	2.88	9000	0.7084	41.9856	0.6531	46.4559
2.7690	3.04	9500	0.7060	42.0913	0.6545	46.4582
2.7337	3.2	10000	0.7045	42.0669	0.6538	46.3729
2.7422	3.36	10500	0.7030	42.0979	0.6539	46.3751
2.7510	3.52	11000	0.7003	42.1584	0.6547	46.3255
2.7949	3.68	11500	0.6992	42.1897	0.6549	46.4093
2.7022	3.84	12000	0.6964	42.2318	0.6550	46.3330
2.7273	4.0	12500	0.6941	42.2760	0.6550	46.2868
2.6612	4.16	13000	0.6955	42.2284	0.6550	46.3091
2.6800	4.32	13500	0.6946	42.2476	0.6550	46.1855
2.6671	4.48	14000	0.6929	42.2365	0.6546	46.2724
2.6884	4.64	14500	0.6924	42.3371	0.6552	46.2344
2.6564	4.8	15000	0.6922	42.3021	0.6550	46.2341
2.6617	4.96	15500	0.6925	42.3001	0.6550	46.2511

ndhieu1101/en-es-marianmt

en-es-marianmt