qtros / carvana-challenge

My repository for the Carvana Image Masking Challenge

Carvana challenge

This repository holds the code for the carnava image making challenge. It's meant to show how to construct Unets with Pytorch in a concise and straightforward way.

Dependencies

• Pytorch 0.2.0
• Sklearn
• Numpy
• OpenCV-Python
• Pillow SIMD (for speedup) or the original Pillow
• Kaggle data downloader
• Tensorboard Pytorch

Usage

In the notebook/ folder you'll find a jupyter notebook containing a little exploratory data analysis. To run the script you'll need to set 2 variables in your environment, KAGGLE_USER and KAGGLE_PASSWD:

export KAGGLE_USER="your_kaggle_username"
export KAGGLE_PASSWD="your_kaggle_password"

This will allow you to automatically check and download the required dataset from Kaggle. When it's done simply execute the main file with:

python src/main.py

If you want to take a look at the prediction at each epochs you can use tensorboard with:

tensorboard --logdir=./logs