jialicheng/samsum_t5-large

samsum_42

This model is a fine-tuned version of google/t5-v1_1-large on the samsum dataset. It achieves the following results on the evaluation set:

• Loss: 1.4591
• Rouge1: 50.8056
• Rouge2: 27.2825
• Rougel: 42.3835
• Rougelsum: 47.2015
• Gen Len: 26.3814

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: 16
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	921	1.7107	49.4859	25.6708	40.6695	45.4657	29.0733
3.053	2.0	1842	1.4591	50.8056	27.2825	42.3835	47.2015	26.3814
1.8078	3.0	2763	1.4260	47.4424	24.5737	38.5731	43.77	24.5098
1.6115	4.0	3684	1.4194	47.6099	24.4291	38.6691	43.651	24.7213
1.5189	5.0	4605	1.4182	47.9053	25.0788	39.0674	43.9306	25.1626

Framework versions

• Transformers 4.39.3
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.15.2