t5-small-finetuned-feedback

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

• Loss: 1.6145
• Rouge1: 51.2809
• Rouge2: 27.3229
• Rougel: 49.2287
• Rougelsum: 49.211
• Gen Len: 10.1736

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: 12
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	61	2.9832	24.9931	10.0881	21.9651	22.0687	16.4876
No log	2.0	122	2.1822	36.3348	17.5969	34.3034	34.2834	12.1653
No log	3.0	183	1.9607	43.7295	21.5907	41.8815	41.929	10.5372
No log	4.0	244	1.8412	48.7074	25.1744	46.8382	46.8399	10.405
No log	5.0	305	1.7674	50.1972	26.4116	48.1456	48.0538	10.2066
No log	6.0	366	1.7195	51.0984	27.8685	48.9483	49.0108	10.3554
No log	7.0	427	1.6832	50.272	27.3168	48.4083	48.4307	10.0331
No log	8.0	488	1.6558	50.6829	27.5132	48.6684	48.735	10.2727
2.363	9.0	549	1.6357	50.0286	27.0674	48.0211	48.0783	10.1736
2.363	10.0	610	1.6240	50.8207	26.8345	48.6528	48.6903	10.1983
2.363	11.0	671	1.6166	50.9796	27.0236	48.8888	48.8958	10.1901
2.363	12.0	732	1.6145	51.2809	27.3229	49.2287	49.211	10.1736

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1