Pegasus_xsum_samsum

This model is a fine-tuned version of google/pegasus-xsum on the samsum dataset. It achieves the following results on the evaluation set:

• Loss: 1.4709
• Rouge1: 0.5072
• Rouge2: 0.2631
• Rougel: 0.4243
• Rougelsum: 0.4244
• Gen Len: 19.1479
• Precision: 0.9247
• Recall: 0.9099
• F1: 0.917

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len	Precision	Recall	F1
1.9542	1.0	920	1.5350	0.4928	0.2436	0.4085	0.4086	18.5672	0.9229	0.9074	0.9149
1.6331	2.0	1841	1.4914	0.5037	0.257	0.4202	0.4206	18.8154	0.9246	0.9092	0.9166
1.5694	3.0	2762	1.4761	0.5071	0.259	0.4212	0.4214	19.4487	0.9241	0.9103	0.917
1.5374	4.0	3680	1.4709	0.5072	0.2631	0.4243	0.4244	19.1479	0.9247	0.9099	0.917

Framework versions

• Transformers 4.36.0
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.15.0