https://github.com/srowhani/comp4107-project
https://github.com/srowhani/comp4107-project
Last synced: about 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/srowhani/comp4107-project
- Owner: srowhani
- Created: 2017-12-14T01:50:57.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2017-12-22T03:03:15.000Z (over 8 years ago)
- Last Synced: 2025-12-29T12:24:21.351Z (6 months ago)
- Language: Jupyter Notebook
- Size: 13.4 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
## Description
Our project compares several variants of convolutional neural networks to
evaluate classification accuracy relative to configuration. We look at several
architectures, employing different tactics to achieve better results.
## Group Member
Seena Rowhani - 100 945 353
Chris Ermel - 100 934 583
## Dependencies
```
import warnings
import _pickle as pickle
from __future__ import division
from sklearn.model_selection import train_test_split
import numpy as np
import tensorflow as tf
from math import sqrt
import matplotlib.pyplot as plt
from scipy.misc import toimage
```
Note - `tensorflow-gpu` was used.
## Testing
Each network can be tested through `jupyter`, or by running the python file.
Classification accuracy can be viewed on tensorboard post-completion - and Estimator
model checkpoints are computed along the way.