learning_rate: 5e-05
train_batch_size: 16
eval_batch_size: 32
seed: 42
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
lr_scheduler_warmup_steps: 0.1
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Bleu
No log	1.0	442	1.7459	0.0058	5.2288
2.8452	2.0	884	1.3527	0.0058	10.0602
1.4873	3.0	1326	1.2063	0.0058	13.7917
1.1715	4.0	1768	1.1297	0.0058	16.0272
0.9706	5.0	2210	1.0952	0.0058	17.1644
0.8288	6.0	2652	1.0784	0.0058	18.7661
0.7138	7.0	3094	1.0871	0.0058	18.8073
0.6201	8.0	3536	1.0806	0.0058	19.8380
0.6201	9.0	3978	1.0916	0.0058	20.1793
0.5580	10.0	4420	1.0985	0.0058	20.0732

danyanderson/m2m100-yemba-french-fr2ybb

m2m100-yemba-french-fr2ybb