This repository compiles different exercises for scholars of the Deep Learning nanodegree from Udacity that want to go the extra mile. This material is not required for Udacity DL phases completion, and will not be counted as an extra in their consideration.
Diving further into the different topics of the course and notebook exercises from the nanodegree repo, you can find the following exercises for each lesson.
Create the building blocks for your own MLP in numpy by completing two challenges:
- Lesson 4 (PyTorch)
- Lesson 5 (Convolutional Neural Network)
- Lesson 6 (Style Transfer)
- Lesson 7 (Recurrent Neural Network)
- Lesson 8 (Sentiment Analysis)
The numpy, matplotlib, torch and torchvision packages are required to properly use the repo. Tested on the following version:
import sys
import numpy, matplotlib, torch, torchvision
print('Python %s' % '.'.join(map(str, sys.version_info[:3])))
print(f'Numpy {numpy.__version__}, Matplotlib {matplotlib.__version__}, PyTorch {torch.__version__}, Torchvision {torchvision.__version__}')
Python 3.6.5
Numpy 1.15.4, Matplotlib 2.2.3, PyTorch 1.0.0, Torchvision 0.2.1
Please refer to PyTorch/Torchvision installation instructions if you haven't installed them yet.
Each lesson has a folder with a folders and other modules. The notebook will contain all the instructions. Fork this repo to complete the exercises, and please notify the author of potential bugs/issues
Run your jupyter notebook server in the same environment you installed the previously mentioned requirements. If you are using Anaconda distribution, open your terminal (Linux/MacOS) or Anaconda prompt (Windows) and run:
jupyter notebook
Now navigate to the corresponding folder and follow the instructions of the notebook.
If you wish to submit a request or report an issue, go to the Issues section, and create a "New issue".
Regarding issues, use the following format for the title:
[Lesson #] Your Issue name
Example: [Lesson 2] Adding more instructions for the layer exercise and these guidelines for the comment:
- Ensure you already have restart your kernel before reporting an issue
- Specify your setup (OS, OS version, Python version, requirements' versions)
- Format and specify the part of the code that is an issue
- Format and specify the error/issue that you encounter
- If relevant, explain what you have already tried to resolve the issue
Regarding requests, use the following title format:
[Request] Your request name
Example: [Request] Creating an exercise illustrating dropout
- Lesson 2
- Lesson 4
- Lesson 5
- Lesson 6
- Lesson 7
- Lesson 8