中文

Flan-T5 (small) fine-tuned on OpenAI summarize_from_feedback for summarizing

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

  • Loss: 2.1488
  • Rouge1: 27.2966
  • Rouge2: 9.5886
  • Rougel: 22.1999
  • Rougelsum: 23.6317
  • Gen Len: 18.9310

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: 32
  • eval_batch_size: 64
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 6

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum Gen Len
2.2472 1.0 2902 2.1882 26.2033 8.83 21.3673 22.7758 18.9234
2.1142 2.0 5804 2.1608 27.1972 9.4269 22.1761 23.6252 18.8796
2.0484 3.0 8706 2.1524 27.0963 9.4578 21.9866 23.5124 18.9033
2.0055 4.0 11608 2.1519 27.2428 9.5514 22.1542 23.6036 18.9347
1.9647 5.0 14510 2.1488 27.2966 9.5886 22.1999 23.6317 18.9310
1.9547 6.0 17412 2.1488 27.5602 9.673 22.3768 23.8399 18.9236

Framework versions

  • Transformers 4.25.1
  • Pytorch 1.13.0+cu116
  • Datasets 2.8.0
  • Tokenizers 0.13.2