domenicrosati's picture
Update README.md
bc93f76
metadata
license: mit
tags:
  - text-classification
  - generated_from_trainer
datasets:
  - paws
metrics:
  - f1
  - precision
  - recall
model-index:
  - name: deberta-v3-large-finetuned-paws-paraphrase-detector
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: paws
          type: paws
          args: labeled_final
        metrics:
          - name: F1
            type: f1
            value: 0.9426698284279537
          - name: Precision
            type: precision
            value: 0.9300853289292595
          - name: Recall
            type: recall
            value: 0.9555995475113123

deberta-v3-large-finetuned-paws-paraphrase-detector

Feel free to use for paraphrase detection tasks!

This model is a fine-tuned version of microsoft/deberta-v3-large on the paws dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3046
  • F1: 0.9427
  • Precision: 0.9301
  • Recall: 0.9556

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: 6e-06
  • 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
  • lr_scheduler_warmup_steps: 50
  • num_epochs: 3
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss F1 Precision Recall
0.1492 1.0 6176 0.1650 0.9537 0.9385 0.9695
0.1018 2.0 12352 0.1968 0.9544 0.9427 0.9664
0.0482 3.0 18528 0.2419 0.9521 0.9388 0.9658

Framework versions

  • Transformers 4.20.1
  • Pytorch 1.11.0
  • Datasets 2.1.0
  • Tokenizers 0.12.1