long-t5-base-govreport

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

long-t5-base-govreport

This model is a fine-tuned version of google/long-t5-tglobal-base on the None dataset. It achieves the following results on the evaluation set:

• Gen Len: 787.34
• Loss: 1.5448
• Rouge1: 57.2303
• Rouge2: 24.9705
• Rougel: 26.8081
• Rougelsum: 54.2747

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

Refer to the pszemraj/govreport-summarization-8192 dataset.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 3
• eval_batch_size: 1
• seed: 4299
• gradient_accumulation_steps: 128
• total_train_batch_size: 384
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 25.0

Training results

Framework versions

• Transformers 4.25.0.dev0
• Pytorch 1.13.0+cu117
• Datasets 2.7.0
• Tokenizers 0.13.2