YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

---language: en-license: mit tags:- distilled- text-classification datasets:- custommetrics:- accuracy- f1- inference-time---

Distilled Model from BERT

This model is a knowledge distilled version of a fine-tuned BERT model.

Model description

  • Distillation parameters: alpha=0.5, temperature=2.0
  • Student model base: DistilBertForSequenceClassification

Performance Metrics

Model Size

  • Teacher model: 416.43 MB
  • Student model: 254.19 MB
  • Size reduction: 38.96%
  • Size ratio: 1.64x

Test Set 1 Performance

  • Teacher accuracy: 0.8459
  • Student accuracy: 0.7945
  • Accuracy difference: -0.0514
  • Teacher F1: 0.8344
  • Student F1: 0.7728
  • Speed improvement: 1.99x

Test Set 2 Performance

  • Teacher accuracy: 0.6710
  • Student accuracy: 0.6217
  • Accuracy difference: -0.0493
  • Teacher F1: 0.6571
  • Student F1: 0.5998
  • Speed improvement: 1.99x

Detailed Metrics - Test Set 1

  • Precision: 0.7771
  • Recall: 0.7945
  • MCC: 0.3811
  • Inference time per sample: 3.52 ms
  • Samples per second: 284.23
Downloads last month
3
Safetensors
Model size
66.6M params
Tensor type
F32
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.