Gradient Descent

Gradient Descent Using Pure Python Without NumPy or SciPy

A progressive guide on how to develop gradient descent code for a given model using python.

This code accompanies the blog post Gradient Descent Using Pure Python without Numpy or Scipy.

Getting started

1. This code accompanies the blog post Gradient Descent Using Pure Python without Numpy or Scipy, and it is best to walk through this code from there.
2. GradDesc1.py is some VERY simple code to show how to solve for line parameters using basic gradient descent techniques.
3. GradDesc2.py takes the next step from GradDesc1.py and divides the necessary tasks into functions.
4. Gradient_Descent_Solver.py holds the creation of a class structure for what was developed in GradDesc2.py.
5. Testing_#.py are test files that import the class from Gradient_Descent_Solver.py. The most interesting of these files is Testing_4.py, which uses the class to accomplish a gradient descent solution on some reasonably noisy fake training data.

Deploying / Publishing

I would appreciate it, as you share your work leveraged from this set of scripts, if you would please keep a referral back to my github repo. Thanks! I'd do the same for you. :-)

Contributing

I am open to share the development and improvements of this with others, but it has been solo up until now. Let me know if you'd like to contribute.

Links

This work came from My GitHub Repo originally.

Licensing

"The code in this project is licensed under MIT license."