gergopool / asura_task1

An example solution for local ML task

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Asura ML Task

An example PyTorch solution for the introductory ML task.

Environment

Make sure you have everything installed from requirements.txt.

Download data

Install your csv and txt files under resources/ folder. If your folder already exists somewhere else, don't worry, you can specify your custom path as the argparse suggest. For more info, run ./download --help

In order to start downloading on multiple threads, run

./download --n-workers 4

Train / Val / Test split

Split your data into three parts.

./split --folder resources/images

Training

Example run:

./train --folder resources/images --model tiny

For more parameters, run ./train --help

About

An example solution for local ML task

License:MIT License


Languages

Language:Python 100.0%