CodeIsAbstract's picture
Model save
8136f45 verified
|
raw
history blame
2.37 kB
metadata
tags:
  - generated_from_trainer
metrics:
  - rouge
model-index:
  - name: denoice-finetuned-xsum
    results: []

denoice-finetuned-xsum

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

  • Loss: 0.6757
  • Rouge1: 73.1568
  • Rouge2: 56.3431
  • Rougel: 73.2739
  • Rougelsum: 73.2387
  • Gen Len: 16.5471

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: 500
  • eval_batch_size: 500
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum Gen Len
No log 1.0 76 0.7087 72.4084 55.2193 72.4899 72.4272 16.5916
No log 2.0 152 0.6998 72.7389 55.4449 72.7601 72.7258 16.5497
No log 3.0 228 0.6946 72.674 55.5275 72.7467 72.712 16.5288
No log 4.0 304 0.6888 72.7071 55.7658 72.7673 72.7402 16.5524
No log 5.0 380 0.6829 72.8829 55.8072 72.9415 72.9187 16.5602
No log 6.0 456 0.6801 73.067 55.9923 73.137 73.1117 16.5681
0.8082 7.0 532 0.6791 73.1192 56.0297 73.2107 73.1619 16.5707
0.8082 8.0 608 0.6768 73.0697 56.0297 73.1433 73.1279 16.5785
0.8082 9.0 684 0.6763 72.9717 55.9654 73.0873 73.0365 16.5576
0.8082 10.0 760 0.6757 73.1568 56.3431 73.2739 73.2387 16.5471

Framework versions

  • Transformers 4.36.2
  • Pytorch 1.13.1
  • Datasets 2.16.1
  • Tokenizers 0.15.0