Edit model card

bert-finetuned-bpmn

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

  • Loss: 0.3456
  • Precision: 0.8113
  • Recall: 0.86
  • F1: 0.8350
  • Accuracy: 0.9341

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: 8
  • eval_batch_size: 8
  • 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 Precision Recall F1 Accuracy
No log 1.0 10 0.2716 0.7778 0.84 0.8077 0.9115
No log 2.0 20 0.2428 0.7669 0.8333 0.7987 0.9160
No log 3.0 30 0.2726 0.7875 0.84 0.8129 0.9205
No log 4.0 40 0.2658 0.7862 0.8333 0.8091 0.9214
No log 5.0 50 0.2470 0.7914 0.86 0.8243 0.9268
No log 6.0 60 0.2745 0.7791 0.8467 0.8115 0.9250
No log 7.0 70 0.3415 0.8280 0.8667 0.8469 0.9259
No log 8.0 80 0.3524 0.775 0.8267 0.8000 0.9178
No log 9.0 90 0.3307 0.8313 0.8867 0.8581 0.9322
No log 10.0 100 0.3161 0.7778 0.84 0.8077 0.9214
No log 11.0 110 0.3646 0.8387 0.8667 0.8525 0.9322
No log 12.0 120 0.3262 0.7925 0.84 0.8155 0.9223
No log 13.0 130 0.3436 0.8462 0.88 0.8627 0.9350
No log 14.0 140 0.3427 0.8516 0.88 0.8656 0.9377
No log 15.0 150 0.3163 0.7950 0.8533 0.8232 0.9322
No log 16.0 160 0.3233 0.8291 0.8733 0.8506 0.9377
No log 17.0 170 0.3354 0.8050 0.8533 0.8285 0.9322
No log 18.0 180 0.3468 0.8291 0.8733 0.8506 0.9341
No log 19.0 190 0.3457 0.8176 0.8667 0.8414 0.9341
No log 20.0 200 0.3456 0.8113 0.86 0.8350 0.9341

Framework versions

  • Transformers 4.30.1
  • Pytorch 2.0.1+cu118
  • Datasets 2.12.0
  • Tokenizers 0.13.3
Downloads last month
13
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.