learning_rate: 0.0003
train_batch_size: 8
eval_batch_size: 8
seed: 42
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: 2
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
3.6484	0.2786	10000	2.2900	0.3932	0.1633	0.251	0.3637	77.2435
2.7934	0.5573	20000	2.1941	0.3971	0.1677	0.2597	0.368	74.5763
2.6941	0.8359	30000	2.1688	0.3996	0.1699	0.2622	0.3702	75.5441
2.654	1.1145	40000	2.1533	0.4008	0.1716	0.2642	0.3713	74.3602
2.6302	1.3931	50000	2.1450	0.4016	0.1726	0.2648	0.3723	74.8291
2.6214	1.6718	60000	2.1411	0.4018	0.1725	0.2651	0.3727	75.1359
2.6178	1.9504	70000	2.1378	0.4018	0.173	0.2656	0.3728	74.6267

BKVNP/bart_prefix_finetune

bart_prefix_finetune