Pegasus-Large-Dataset-Psycho

This model is a fine-tuned version of google/pegasus-cnn_dailymail on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.8738
• Rouge1: 0.6063
• Rouge2: 0.3841
• Rougel: 0.5184
• Rougelsum: 0.5186
• Gen Len: 92.0949

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	137	1.1358	0.5445	0.315	0.4514	0.4515	99.438
No log	2.0	274	1.0206	0.5717	0.3415	0.4779	0.4779	93.0292
No log	3.0	411	0.9631	0.5851	0.3577	0.4944	0.4944	92.5109
1.2737	4.0	548	0.9283	0.5918	0.3651	0.501	0.5011	92.4526
1.2737	5.0	685	0.9017	0.5987	0.3764	0.5105	0.5107	91.9781
1.2737	6.0	822	0.8863	0.6032	0.3811	0.5155	0.5156	93.8467
1.2737	7.0	959	0.8767	0.6063	0.3835	0.519	0.519	92.2044
0.9978	8.0	1096	0.8738	0.6063	0.3841	0.5184	0.5186	92.0949

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1