shhossain's picture
Update README.md
b4f47de
|
raw
history blame
2.14 kB
metadata
license: mit
datasets:
  - shhossain/book-text-classifier
language:
  - en
pipeline_tag: text-classification
widget:
  - text: >-
      Shen Yuanye didn’t expect to go back to Huang Ni’s matter again, this
      matter of being killed without finding the murderer, who else could be
      beside her.
    example_title: Book Text
  - text: >-
      I am so sorry this is a day late, guys. Unfortunately, my internet was
      down so it was out of my control. Its still intermittent but hopefully it
      will be fine by next week.
    example_title: Normal Text
metrics:
  - accuracy
model-index:
  - name: shhossain/bert-tiny-book-text-classifier
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          type: shhossain/book-text-classifier
          name: book-text-classifier
          split: test
        metrics:
          - type: accuracy
            value: 0.999128

Book Test Classifier

Classify book text (mostly fictional book)

Model Details

Model Description

This model is finetuned on bert-tiny for classifying book text.

  • Developed by: shhossain
  • Model type: [Bert]
  • Language(s) (NLP): [English]
  • License: [MIT]
  • Finetuned from model [Bert-Tiny]: bert-tiny

Uses

from transformers import pipeline, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained('prajjwal1/bert-tiny')

pipe = pipeline('text-classification', model='shhossain/bert-tiny-book-text-classifier')

book_text = """Shen Yuanye didn’t expect to go back to Huang Ni’s matter again, this matter of being killed without finding the murderer, who else could be beside her."""

pipe(book_text) # LABEL_1
>> [{'label': 'LABEL_1', 'score': 0.9998537302017212}]

normal_text = """I am so sorry this is a day late, guys. Unfortunately, my internet was down so it was out of my control. Its still intermittent but hopefully it will be fine by next week. Hopefully the fact that Skye and Pietro are back in form will help make up for it."""

pipe(normal_text) # LABEL_0
>> [{'label': 'LABEL_0', 'score': 0.9984021782875061}]