pegasus-large_1742608050.119433

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

• Loss: 4.8799
• Rouge1: 46.7629
• Rouge2: 17.2538
• Rougel: 27.775
• Rougelsum: 32.6028
• Gen Len: 85.8611

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	331	5.1130	48.5686	16.6489	28.3029	32.3136	71.2562
5.3363	2.0	662	5.0415	49.2151	17.0482	28.5422	32.4827	70.0833
5.3363	3.0	993	4.9860	48.9739	17.1901	28.4601	32.6516	72.8179
5.0157	4.0	1324	4.9619	48.2866	17.4271	28.1784	32.8011	77.8704
4.9426	5.0	1655	4.9329	48.7697	17.7741	28.6454	33.1256	76.4846
4.9426	6.0	1986	4.9133	48.0678	17.7361	28.3205	32.9328	79.4753
4.863	7.0	2317	4.8973	47.1789	17.4753	27.9637	32.7952	85.8056
4.8272	8.0	2648	4.8786	47.3498	17.2852	27.9143	32.7426	82.9198
4.8272	9.0	2979	4.8805	47.1749	17.4155	27.8944	32.7537	85.1111
4.7888	10.0	3310	4.8799	46.7629	17.2538	27.775	32.6028	85.8611

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0