learning_rate: 3e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 16
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: 500
num_epochs: 5
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Rouge3
11.4824	1.0	117	4.9072	39.2430	20.8989	35.8162	35.8655	8.7146
4.497	2.0	234	1.3738	42.8823	23.7662	39.1299	39.2109	11.1847
1.5252	3.0	351	0.2833	44.9749	26.9850	41.5834	41.7102	13.1019
0.3443	4.0	468	0.2533	45.8865	25.7053	42.1245	42.2211	12.9262
0.2507	5.0	585	0.2390	46.9332	28.3688	43.8828	43.9145	15.6362

Ftmhd/bart-base-finetuned-steel-news-production

bart-base-finetuned-steel-news-production