multinews_model

This model is a fine-tuned version of t5-small on the multi_news dataset. It achieves the following results on the evaluation set:

• Loss: 2.7165
• Rouge1: 0.1482
• Rouge2: 0.0472
• Rougel: 0.1132
• Rougelsum: 0.1132
• Gen Len: 19.0

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 10
• eval_batch_size: 10
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	450	2.8616	0.1388	0.0418	0.1057	0.1056	19.0
3.2544	2.0	900	2.7991	0.1427	0.0438	0.1089	0.1089	19.0
2.999	3.0	1350	2.7693	0.1449	0.046	0.1115	0.1114	19.0
2.958	4.0	1800	2.7531	0.1466	0.0462	0.112	0.1118	19.0
2.9198	5.0	2250	2.7431	0.1466	0.0465	0.112	0.1119	19.0
2.8838	6.0	2700	2.7328	0.1474	0.0461	0.1125	0.1123	19.0
2.8774	7.0	3150	2.7270	0.1477	0.0463	0.1126	0.1124	19.0
2.8712	8.0	3600	2.7226	0.148	0.0466	0.1128	0.1127	19.0
2.854	9.0	4050	2.7197	0.1479	0.047	0.1129	0.1128	19.0
2.8541	10.0	4500	2.7188	0.1485	0.0471	0.113	0.1129	19.0
2.8541	11.0	4950	2.7168	0.1483	0.0472	0.1131	0.1131	19.0
2.8466	12.0	5400	2.7165	0.1482	0.0472	0.1132	0.1132	19.0

Framework versions

• Transformers 4.33.3
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3