Edit model card

synonyms_5000_plus_3000_default_3_epoch

This model is a fine-tuned version of facebook/bart-large on the midas/inspec generation dataset, concatenated with data augmented custom artemfilipenko/synonyms-augmented-5x-inspec dataset. It achieves the following results on the evaluation set:

  • Loss: 1.7956
  • F1@5ext: 0.4590
  • P@5ext: 0.6116
  • R@5ext: 0.4109
  • F1@10ext: 0.5403
  • P@10ext: 0.5953
  • R@10ext: 0.5374
  • F1@5abs: 0.2019
  • P@5abs: 0.3080
  • R@5abs: 0.1721
  • F1@10abs: 0.2307
  • P@10abs: 0.3066
  • R@10abs: 0.2109
  • F1@oext: 0.5427
  • P@oext: 0.6045
  • R@oext: 0.5246
  • F1@oabs: 0.2316
  • P@oabs: 0.3079
  • R@oabs: 0.2094

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: 4
  • eval_batch_size: 10
  • seed: 42
  • gradient_accumulation_steps: 6
  • total_train_batch_size: 24
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 3.0

Training results

Framework versions

  • Transformers 4.24.0.dev0
  • Pytorch 1.9.0+cu111
  • Datasets 2.6.1
  • Tokenizers 0.13.1
Downloads last month
10
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Datasets used to train artemfilipenko/keyphrase-generation-bart-large-trained-on-augmented-and-default-inspec