making better loss history / epoch plot

savving memory with close fig

switching to png

think that the way we had teh plot function setup may have been causing some problems with buffers in the cloud

believe to have created a function to build a svg plot and send it over the network

adding cors support for this api because we are starting to work on our web application

fixing bug with algorithm as an argument for clustering, this was breaking the neural network endpoint

kmeans is ready for deployment

kmeans with example plot done and ready for deployment

moving on from kmeans to kmedoids

kmeans clustering works and returns centroids and labeled data

v1 of kmean may be complete

think I am done with build func

leaving off here, working on kmeans algorithm

working on kmeans

updating gitignore

starting work with clustering algorithms

python example for creating a plot of loss history over time for neural_network

adding jupyter notebook example

algorithm specific documentation in readme

working on making readme better

cleaning up readme, now moving on to building clustering algorithms from scratch

fixed dimensions issue with the iris dataset

having some issues with backprop when trying to use the iris dataset

fixing train test split issue with iris dataset

trying this commit for iris

trying to move forwards with the iris dataset

showing sample output for simple request in readme

updating requirements.txt

updating example request in readme

messing around with readme

seeing if the debug mode is causing my problems

re configuring dockerfile

render did not seem to like my type hints in the logs

added dockerfile

allowing a get request just cause

successfully returning data on how the neural network performed

updating requirements.txt and cleaning things up. Thought it would be working but for some reason requests from postman are not making it to the api

working with serializing the model with to_dict()

produced a solution to the problem of funcions not being accessed

working on debugging parameters inside of neural network

setting it up as a straightforward rest api

cutting the front end mess and just using this as a backend api

splitting up some backprop functions that may have been too long

working on making it just a flask app, got rid of the main.py stuff

getting rid of lambda funcs

configured the run_dev shell script

tailwind working in the app

focus on the math first

ignoring node stuff so we can use tailwind and not ruin repo