t5-arxiv

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.3852
• Rouge1: 18.0722
• Rouge2: 6.8453
• Rougel: 14.3659
• Rougelsum: 16.4137
• Gen Len: 19.0

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: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.5169	1.0	12500	2.3852	18.0722	6.8453	14.3659	16.4137	19.0

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu113
• Datasets 2.2.2
• Tokenizers 0.12.1