bert-base-finetuned-sts-v3

This model is a fine-tuned version of klue/bert-base on the klue dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3716
  • Pearsonr: 0.9172

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: 4
  • eval_batch_size: 4
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 20

Training results

Training Loss Epoch Step Validation Loss Pearsonr
0.2265 1.0 2917 0.4886 0.8933
0.1504 2.0 5834 0.4374 0.8948
0.0982 3.0 8751 0.5246 0.8957
0.0832 4.0 11668 0.4387 0.9006
0.0751 5.0 14585 0.4036 0.9049
0.0564 6.0 17502 0.3828 0.9133
0.0488 7.0 20419 0.3716 0.9172
0.0384 8.0 23336 0.4060 0.9093
0.0365 9.0 26253 0.3939 0.9065
0.0319 10.0 29170 0.3953 0.9106
0.0262 11.0 32087 0.3885 0.9109
0.0219 12.0 35004 0.3724 0.9154
0.0188 13.0 37921 0.3827 0.9111
0.0175 14.0 40838 0.4103 0.9099
0.0144 15.0 43755 0.3768 0.9152
0.0132 16.0 46672 0.3868 0.9151
0.0125 17.0 49589 0.3981 0.9103
0.0106 18.0 52506 0.3808 0.9138
0.0095 19.0 55423 0.3904 0.9128
0.0089 20.0 58340 0.3885 0.9137

Framework versions

  • Transformers 4.25.1
  • Pytorch 1.13.0
  • Datasets 2.7.1
  • Tokenizers 0.13.2
Downloads last month
4
Inference Providers NEW
This model is not currently available via any of the supported third-party Inference Providers, and the model is not deployed on the HF Inference API.

Dataset used to train namvandy/bert-base-finetuned-sts-v3

Evaluation results