Skip to content

Code for GalaxyZoo challenge, Computer Vision course final project.

Notifications You must be signed in to change notification settings

guangylegend/Galaxy-zoo-challenge

Repository files navigation

Guang Yang (gy552) & Chen Zhang (cz1389)

Description

We implemented several models. For your convinence, we placed each banch in a seperate directory.
The directories are name as "GalaxyZoo-"
In each directory,
-- download.sh # A shell script allowing you to download the data used for the task.
# Your need to export your cookie for the Kaggle challenge and put it in cookie.txt.
# Or you can just download the data from the browser.
-- main.py # The skeleton of the training and validation.
-- data.py # Data preparation including preprocessing and augmentation.
-- model.py # The model definition.
-- evaluate.py # Generate the test output.

The files are brought from assignment 3, therefore have similar command line argument.

Usage

To train the model, invoke

python main.py --epochs 200 --model_name <model_name>

It will find the local minimums of the validation loss, and dump the model as _.pth

To generate the test output, invoke

python evaluate.py --model _.pth

About

Code for GalaxyZoo challenge, Computer Vision course final project.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published