t5-small-finetuned-xsum

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

• Loss: 2.5074
• Rouge1: 27.9257
• Rouge2: 7.4618
• Rougel: 21.9338
• Rougelsum: 21.9405
• Gen Len: 18.8176

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 0.5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.817	0.0784	500	2.5683	26.6596	6.6324	20.7701	20.7761	18.8057
2.8029	0.1568	1000	2.5435	27.1558	6.9694	21.2178	21.2216	18.7999
2.7797	0.2352	1500	2.5270	27.5528	7.2608	21.621	21.6233	18.7982
2.7651	0.3137	2000	2.5165	27.6104	7.2896	21.6928	21.7012	18.8133
2.7514	0.3921	2500	2.5112	27.8452	7.3791	21.8632	21.8659	18.8118
2.7463	0.4705	3000	2.5074	27.9257	7.4618	21.9338	21.9405	18.8176

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1