billsum_236_t5-base

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

• Loss: 2.1143
• Rouge1: 0.1513
• Rouge2: 0.0546
• Rougel: 0.1244
• Rougelsum: 0.1245
• Gen Len: 18.979

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
1.2695	1.69	500	2.1398	0.157	0.0611	0.1279	0.1279	19.0
1.1033	3.38	1000	2.1182	0.1582	0.0629	0.1296	0.1297	18.9984
1.1178	5.07	1500	2.1133	0.1551	0.0594	0.1275	0.1277	18.979
1.0399	6.75	2000	2.1171	0.1538	0.058	0.1266	0.1266	18.9887
1.0364	8.44	2500	2.1143	0.1513	0.0546	0.1244	0.1245	18.979

Framework versions

• Transformers 4.36.2
• Pytorch 2.0.1+cu117
• Datasets 2.15.0
• Tokenizers 0.15.0