based-bert-sc / README.md
0xMaka's picture
[+]
27fac14
|
raw
history blame
1.16 kB
metadata
datasets:
  - 0xMaka/trading-candles-subset-sc-format
language:
  - en
metrics:
  - accuracy
  - f1
widget:
  - text: 'identify candle: 17284.58,17264.41,17284.58,17264.41'
    example_title: Bear
  - text: >-
      identify candle: open: 17343.43, close: 17625.18, high: 17804.68, low:
      17322.15
    example_title: Bull
license: gpl

Based Bert for sequence classification

This model is a POC and shouldn't be used for any production task.

Model description

Based Bert SC is a text classification bot for binary classification of a trading candles opening and closing prices.

Uses and limitations

This model can reliably return the bullish or bearish status of a candle given the opening and closing, high and low in a format shown. It will have trouble if the order of the numbers change (even if tags are included).

How to use

You can use this model directly with a pipeline

>>> from transformers import pipeline
>>> pipe = pipeline("text-classification", model="0xMaka/based-bert-sc")
>>> text = "identify candle: open: 21788.19, close: 21900, high: 21965.23, low: 21788.19"
>>> pipe(text)
[{'label': 'Bullish', 'score': 0.9999682903289795}]