SuperTML

Author: Ioannis Gatopoulos, 2020

Description

PyTorch implementation of research paper:

SuperTML: Two-Dimensional Word Embedding for the Precognition on Structured Tabular Data

Abstrack: Projects tabular data to images creating 2-dimentional embenddings. Then, with the use of a pretrained model, it is capable of performing regression and classification tasks.

• Code • Paper • Blog-Post

Results

Required Python packages

Install all the dependencies:

pip3 install -r requirements.txt

Run

cd super_tml
python main.py --dataset iris --model densenet121

References

- Baohua Sun.
SuperTML: Two-Dimensional Word Embedding for the Precognition on Structured Tabular Data.
CVPR Workshop Paper, 2019.