pegasus-samsum

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

• Loss: 1.3775

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: 5e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 16
• optimizer: Use 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: 100
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
2.0058	0.1086	100	1.6564
1.6976	0.2172	200	1.5318
1.6041	0.3258	300	1.4837
1.6314	0.4344	400	1.4586
1.6014	0.5430	500	1.4401
1.6482	0.6516	600	1.4302
1.6023	0.7602	700	1.4151
1.621	0.8689	800	1.4111
1.5236	0.9775	900	1.4033
1.4723	1.0858	1000	1.4004
1.4854	1.1944	1100	1.3961
1.4423	1.3030	1200	1.3924
1.5885	1.4116	1300	1.3860
1.4873	1.5202	1400	1.3853
1.4214	1.6288	1500	1.3801
1.3327	1.7374	1600	1.3794
1.4677	1.8460	1700	1.3781
1.4413	1.9547	1800	1.3775

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0