pegasus-newsroom-rewriter

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

• Loss: 1.3424
• Rouge1: 46.6856
• Rouge2: 31.6377
• Rougel: 33.2741
• Rougelsum: 44.5003
• Gen Len: 126.58

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: 1
• eval_batch_size: 1
• seed: 42
• 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
No log	1.0	450	1.4020	47.0593	32.2065	33.9168	44.901	126.32
1.9944	2.0	900	1.3567	46.2635	30.9959	32.933	44.1659	126.48
1.6511	3.0	1350	1.3449	46.1544	30.7257	32.693	43.9977	126.4
1.5951	4.0	1800	1.3424	46.6856	31.6377	33.2741	44.5003	126.58

Framework versions

• Transformers 4.17.0
• Pytorch 1.10.0+cu111
• Datasets 2.0.0
• Tokenizers 0.11.6