roberta-base-finetuned-sts

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

  • Loss: 0.1999
  • Pearsonr: 0.9560

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: 1e-05
  • train_batch_size: 32
  • 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: 200
  • num_epochs: 15
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Pearsonr
No log 1.0 329 0.2462 0.9478
1.2505 2.0 658 0.1671 0.9530
1.2505 3.0 987 0.1890 0.9525
0.133 4.0 1316 0.2360 0.9548
0.0886 5.0 1645 0.2265 0.9528
0.0886 6.0 1974 0.2097 0.9518
0.0687 7.0 2303 0.2281 0.9523
0.0539 8.0 2632 0.2212 0.9542
0.0539 9.0 2961 0.1843 0.9532
0.045 10.0 3290 0.1999 0.9560
0.0378 11.0 3619 0.2357 0.9533
0.0378 12.0 3948 0.2134 0.9541
0.033 13.0 4277 0.2273 0.9540
0.03 14.0 4606 0.2148 0.9533
0.03 15.0 4935 0.2207 0.9534

Framework versions

  • Transformers 4.17.0
  • Pytorch 1.10.0+cu111
  • Datasets 2.0.0
  • Tokenizers 0.11.6
Downloads last month
44
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.

Dataset used to train rurupang/roberta-base-finetuned-sts

Evaluation results