Edit model card

summarise_v4

This model is a fine-tuned version of allenai/led-base-16384 on the multi_news dataset. It achieves the following results on the evaluation set:

  • Loss: 2.5264
  • Rouge2 Precision: 0.1349
  • Rouge2 Recall: 0.1187
  • Rouge2 Fmeasure: 0.1227

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: 2
  • eval_batch_size: 2
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 1
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Rouge2 Precision Rouge2 Recall Rouge2 Fmeasure
2.9616 0.08 10 2.8008 0.0552 0.1944 0.0844
2.7112 0.16 20 2.7017 0.1099 0.1212 0.1078
2.6842 0.24 30 2.6653 0.119 0.1252 0.1157
2.4638 0.32 40 2.6306 0.1386 0.1153 0.1222
2.646 0.4 50 2.6099 0.1449 0.1095 0.122
2.5128 0.48 60 2.5945 0.1259 0.1484 0.1313
2.6737 0.56 70 2.5832 0.1192 0.1252 0.118
2.614 0.64 80 2.5616 0.1288 0.1179 0.1193
2.4643 0.72 90 2.5612 0.1371 0.1227 0.124
2.3164 0.8 100 2.5606 0.1372 0.1177 0.1223
2.4514 0.88 110 2.5339 0.1412 0.1276 0.128
2.8113 0.96 120 2.5264 0.1349 0.1187 0.1227

Framework versions

  • Transformers 4.21.3
  • Pytorch 1.12.1+cu113
  • Datasets 2.6.2.dev0
  • Tokenizers 0.12.1
Downloads last month
4

Dataset used to train debbiesoon/summarise_v4