davidgaofc/SFT_Med_t

results2

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

• Loss: 1.2266
• Rouge1: 0.2682
• Rouge2: 0.1194
• Rougel: 0.2208
• Rougelsum: 0.221
• Gen Len: 154.5226

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
2.6515	1.0	718	1.5195	0.0099	0.0025	0.0087	0.0086	11.3189
1.731	2.0	1436	1.3920	0.1689	0.05	0.1367	0.1366	138.9988
1.5415	3.0	2154	1.3232	0.206	0.0601	0.1642	0.1642	153.786
1.4993	4.0	2872	1.2865	0.2082	0.0622	0.1651	0.1651	151.953
1.4502	5.0	3590	1.2640	0.2366	0.087	0.1883	0.1884	153.7628
1.4226	6.0	4308	1.2491	0.2526	0.1083	0.2053	0.2057	154.0902
1.4175	7.0	5026	1.2385	0.2654	0.1183	0.2168	0.2171	152.6171
1.3855	8.0	5744	1.2319	0.2661	0.118	0.2184	0.2185	153.4085
1.3956	9.0	6462	1.2279	0.2685	0.1194	0.2207	0.2208	154.528
1.3978	10.0	7180	1.2266	0.2682	0.1194	0.2208	0.221	154.5226

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.1+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0