Edit model card

gerskill-bert-job-extended

This model is a fine-tuned version of dathi103/bert-job-german-extended on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0894
  • Hard: {'precision': 0.8381742738589212, 'recall': 0.8898678414096917, 'f1': 0.8632478632478633, 'number': 454}
  • Soft: {'precision': 0.7976190476190477, 'recall': 0.8170731707317073, 'f1': 0.8072289156626505, 'number': 82}
  • Overall Precision: 0.8322
  • Overall Recall: 0.8787
  • Overall F1: 0.8548
  • Overall Accuracy: 0.9776

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: 32
  • eval_batch_size: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 5

Training results

Training Loss Epoch Step Validation Loss Hard Soft Overall Precision Overall Recall Overall F1 Overall Accuracy
No log 1.0 178 0.1001 {'precision': 0.6720720720720721, 'recall': 0.8215859030837004, 'f1': 0.7393458870168483, 'number': 454} {'precision': 0.6931818181818182, 'recall': 0.7439024390243902, 'f1': 0.7176470588235295, 'number': 82} 0.6750 0.8097 0.7362 0.9602
No log 2.0 356 0.0736 {'precision': 0.8148148148148148, 'recall': 0.8237885462555066, 'f1': 0.8192771084337348, 'number': 454} {'precision': 0.7972972972972973, 'recall': 0.7195121951219512, 'f1': 0.7564102564102565, 'number': 82} 0.8124 0.8078 0.8101 0.9744
0.1105 3.0 534 0.0760 {'precision': 0.8280922431865828, 'recall': 0.8700440528634361, 'f1': 0.8485499462943071, 'number': 454} {'precision': 0.8227848101265823, 'recall': 0.7926829268292683, 'f1': 0.8074534161490684, 'number': 82} 0.8273 0.8582 0.8425 0.9768
0.1105 4.0 712 0.0821 {'precision': 0.820040899795501, 'recall': 0.8832599118942731, 'f1': 0.8504772004241782, 'number': 454} {'precision': 0.7790697674418605, 'recall': 0.8170731707317073, 'f1': 0.7976190476190477, 'number': 82} 0.8139 0.8731 0.8425 0.9760
0.1105 5.0 890 0.0894 {'precision': 0.8381742738589212, 'recall': 0.8898678414096917, 'f1': 0.8632478632478633, 'number': 454} {'precision': 0.7976190476190477, 'recall': 0.8170731707317073, 'f1': 0.8072289156626505, 'number': 82} 0.8322 0.8787 0.8548 0.9776

Framework versions

  • Transformers 4.38.1
  • Pytorch 2.1.2+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
7
Safetensors
Model size
112M params
Tensor type
F32
·
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.

Model tree for dathi103/gerskill-bert-job-extended

Finetuned
(1)
this model