Convolutional Neural Networks and image classification

AI Sunsets (Workshop)

Coimbra, September 28th

Software requirements

• Python 3
• Jupyter notebook
• Chrome extension: Fatkun.
• This repository
• The zip files used in the code are also at Google Drive.

Packages

• TensorFlow, Keras
• numpy, pandas, matplotlib
• six.moves (cPickle)

Anaconda

It is recommended that the participants download the Anaconda Distribution to obtain the necessary software.

Once installed, open it and check if all the packages above are installed, or install them otherwise.

Google Colab

Google also provides the platform Colab that allows running the code directly from their servers and requires no set up.

This solution requires an internet connection and the computational power provided may vary.

References

Convolutional Neural Networks in Tensorflow (Coursera)

Siraj Raval (Build a TensorFlow Image Classifier in 5 Min)