t5-small-finetuned-samsum

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

• Loss: 1.7409
• Rouge1: 42.6713
• Rouge2: 19.8452
• Rougel: 35.971
• Rougelsum: 39.6113
• Gen Len: 16.6381

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: 16
• eval_batch_size: 16
• 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
2.2617	1.0	921	1.8712	40.1321	17.123	33.1845	37.13	16.5685
2.0294	2.0	1842	1.8208	41.0756	18.1787	34.4685	38.1966	16.6308
1.9769	3.0	2763	1.7959	41.3228	18.4732	34.6591	38.2431	16.3875
1.9406	4.0	3684	1.7740	41.658	18.7294	34.907	38.6251	16.7078
1.9185	5.0	4605	1.7638	41.8923	19.1845	35.2485	38.7469	16.5428
1.8981	6.0	5526	1.7536	42.3314	19.2761	35.4452	39.3067	16.7579
1.8801	7.0	6447	1.7472	42.362	19.4885	35.7207	39.274	16.5538
1.868	8.0	7368	1.7452	42.3388	19.4036	35.6189	39.2259	16.577
1.8667	9.0	8289	1.7413	42.7453	19.932	36.08	39.7062	16.6736
1.8607	10.0	9210	1.7409	42.6713	19.8452	35.971	39.6113	16.6381

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0