longt5_xl_gov_memsum_40

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

• Loss: 3.0292

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: 0.001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• num_epochs: 15.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.0689	1.0	68	2.9102
0.0658	1.99	136	2.8207
0.0549	2.99	204	2.8665
0.0552	3.99	272	2.8846
0.0534	5.0	341	2.7891
0.0465	6.0	409	3.0601
0.0494	6.99	477	2.9361
0.0477	7.99	545	2.9819
0.0463	8.99	613	3.0402
0.0642	10.0	682	2.9308
0.0601	10.99	750	2.8829
0.0577	11.99	818	2.8970
0.0529	12.99	886	2.9596
0.051	14.0	955	2.9287
0.0514	14.95	1020	3.0292

Framework versions

• Transformers 4.38.1
• Pytorch 2.2.1+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2