gayanin's picture
update model card README.md
a9b9e2a
metadata
license: apache-2.0
tags:
  - generated_from_trainer
model-index:
  - name: t5-small-paraphrasing-mlm
    results: []

t5-small-paraphrasing-mlm

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

  • Loss: 0.7030
  • Rouge2 Precision: 0.6576
  • Rouge2 Recall: 0.4712
  • Rouge2 Fmeasure: 0.532

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: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Rouge2 Precision Rouge2 Recall Rouge2 Fmeasure
0.9215 1.0 13833 0.8050 0.6352 0.454 0.5131
0.855 2.0 27666 0.7679 0.6411 0.4589 0.5184
0.8387 3.0 41499 0.7464 0.6464 0.4626 0.5226
0.8267 4.0 55332 0.7315 0.6513 0.4671 0.5273
0.7879 5.0 69165 0.7217 0.6534 0.4687 0.529
0.7738 6.0 82998 0.7142 0.6548 0.4688 0.5295
0.7793 7.0 96831 0.7094 0.6553 0.4694 0.53
0.7654 8.0 110664 0.7056 0.6573 0.4704 0.5313
0.7675 9.0 124497 0.7036 0.6577 0.4712 0.532
0.7662 10.0 138330 0.7030 0.6576 0.4712 0.532

Framework versions

  • Transformers 4.17.0
  • Pytorch 1.10.0+cu111
  • Datasets 1.18.4
  • Tokenizers 0.11.6