learning_rate: 1e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
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_ratio: 0.06
num_epochs: 20

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
4.6006	1.56	20	5.0770	0.2559	0.1268	0.253	0.2529
2.9074	3.08	40	5.1504	0.2786	0.1408	0.2753	0.2752
2.2547	4.64	60	5.2866	0.2945	0.1433	0.289	0.2893
3.2227	6.16	80	5.9278	0.31	0.1577	0.3039	0.3036
1.7171	7.72	100	5.9767	0.3046	0.1471	0.2977	0.2978
1.5884	9.24	120	6.1892	0.2943	0.1416	0.2893	0.289
1.3678	10.8	140	6.4147	0.2908	0.1432	0.2844	0.2845
1.181	12.32	160	6.4444	0.2971	0.1476	0.2894	0.2892
1.0218	13.88	180	6.7040	0.2911	0.1423	0.2842	0.2835

taiypeo/bart-large-aeslc-rouge-3-loss-differentiable-100-cnt-supervised

bart-large-aeslc-rouge-3-loss-differentiable-100-cnt-supervised