ft-t5-with-dill-sum / README.md
aisuko's picture
End of training
4e0a9c9 verified
metadata
license: apache-2.0
base_model: t5-small
tags:
  - generated_from_trainer
datasets:
  - bills-summarization
metrics:
  - rouge
model-index:
  - name: ft-t5-with-dill-sum
    results:
      - task:
          name: Summarization
          type: summarization
        dataset:
          name: billsum
          type: bills-summarization
        metrics:
          - name: Rouge1
            type: rouge
            value: 0.1441

ft-t5-with-dill-sum

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

  • Loss: 2.5658
  • Rouge1: 0.1441
  • Rouge2: 0.0526
  • Rougel: 0.1184
  • Rougelsum: 0.1184
  • 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: 32
  • eval_batch_size: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 5
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum Gen Len
No log 1.0 31 3.0881 0.1395 0.0436 0.1155 0.1154 19.0
No log 2.0 62 2.7574 0.1315 0.0394 0.109 0.1093 19.0
No log 3.0 93 2.6284 0.1377 0.0465 0.1138 0.1139 19.0
No log 4.0 124 2.5796 0.1428 0.0512 0.1177 0.1175 19.0
No log 5.0 155 2.5658 0.1441 0.0526 0.1184 0.1184 19.0

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.0.0
  • Datasets 2.15.0
  • Tokenizers 0.15.0