Tai-Chi Engine

Powerful deep learning for civilians

太极引擎 深度学习： 强大、多模态、灵活、平民化

See Documentation 看文档

Essence of the Tai-Chi Engine

• Close to state-of-the-art Deep learning, friendly to office folks, all coding-free, clicks away.
• Flexible, supporting multiple kinds of data (image, text, category, multi-category, etc), multiple x at the same time.
• Which columns to be x? Which column to be y? You can decide and play, see if the AI finds out how to guess.

Our big pitch 🎁

• If you're a coding muggle - play with the engine, you'll understand ideas around Deep Learning, and have good model.
• If you're a pro - it's still wildly fun to try new models in 2 minutes' click, especially you have around a dozen columns.

Playing Tai-Chi Engine

First, tell me you are already in a jupyter notebook environment.

Or just using the free kaggle kernel, example here

Or try our colab tutorial:

Open up a table, a pandas dataframe, from excel or from csv file or from SQL database, dosn't matter.

import pandas as pd
df = pd.read_csv('your_data.csv')

Then, you can use the following code to play with the engine.

from tai_chi_engine import TaiChiEngine
# load the engine
engine = TaiChiEngine(df, project="./where/to_save/your_model")
# start the playing
engine()

Good to go! 🎸

Installation 📦

Default installation:

pip install tai-chi-engine

Run App on Trained Model 🚀

You can build prototype App (Based on streamlit) based on trained project, see app part of this library.

ALL kinds of projects, only one way and only few lines to start the app.

Links 🪁

• The github repository is here

For Developers:

• How frontend part of the engine works? See here