Udacity Deep Laerning Foundations Nanodegree
Project 2: Image classification
In this project, I will classify images of the set CIFAR-10 link. This set contains images of airplanes, dogs, cats and other objects. The data needs to be preprocessed, then a convolutional neural network must be trained. It is necessary to normalize the images, make the one-hot encoding of the labels, and assemble the convolutional layers, max pooling and fully connected . At the end, the classifications proposed by the neural network for the images will be presented.
Running using conda!
Run this in command line
Step 1: Create a new environment
conda create --name image-classification python=3
Step 2: Use the new env
source activate image-classification
Step 3: Install dependencies
conda install -c conda-forge tensorflow=1.0.0 conda install -c conda-forge tqdm=4.11.2 conda install matplotlib scikit-learn jupyter notebook
Step 4: Open the notebook to run it
jupyter notebook dlnd_image_classification.ipynb
This folder contains files for Udacity Deep Laerning Foundations Nanodegree Project 2: Image Classification.
dlnd_image_classification.ipynb - Main project file.
dlnd_image_classification.html - Neural network prediction results file.
problem_unittests.py - Unit tests provided by Udacity.
helper.py - Help functions provided by Udacity.