Hartunka's picture
End of training
f202948 verified
metadata
library_name: transformers
language:
  - en
base_model: Hartunka/tiny_bert_km_20_v2
tags:
  - generated_from_trainer
datasets:
  - glue
metrics:
  - accuracy
  - f1
model-index:
  - name: tiny_bert_km_20_v2_qqp
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: GLUE QQP
          type: glue
          args: qqp
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.8167697254513975
          - name: F1
            type: f1
            value: 0.7368757547772963

tiny_bert_km_20_v2_qqp

This model is a fine-tuned version of Hartunka/tiny_bert_km_20_v2 on the GLUE QQP dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4227
  • Accuracy: 0.8168
  • F1: 0.7369
  • Combined Score: 0.7768

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: 256
  • eval_batch_size: 256
  • seed: 10
  • optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Accuracy F1 Combined Score
0.497 1.0 1422 0.4556 0.7813 0.6727 0.7270
0.4091 2.0 2844 0.4252 0.8003 0.7087 0.7545
0.3508 3.0 4266 0.4284 0.8118 0.7313 0.7716
0.3022 4.0 5688 0.4227 0.8168 0.7369 0.7768
0.2622 5.0 7110 0.4398 0.8212 0.7536 0.7874
0.2268 6.0 8532 0.4587 0.8229 0.7595 0.7912
0.2005 7.0 9954 0.4909 0.8228 0.7602 0.7915
0.1759 8.0 11376 0.5125 0.8242 0.7618 0.7930
0.1564 9.0 12798 0.5787 0.8236 0.7680 0.7958

Framework versions

  • Transformers 4.50.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.21.1