I offer these programs for free with a few caveats:

• none of these programs are production ready
• there are no error checking, fool proofing in the code
• I wrote these code for myself to learn how the algorithms work - the nuts and bolts stuff - there may be logical errors
• all of these programs I tested only in a Chrome browser
• these are my "notes", tests, nothing more, but you are welcmoe to use it

[ divisibility_test_perceptron.html ]
[ gradnt_desc_perceptron.html ]
[ linr_regrs.html ]
[ linr_regrs_x.html ]
[ matrix_generator.html ]
[ neural_net2.html ]
[ neural_network_imperial_college2c.html ]
[ neural_network_titanic5b.html ]
[ neural_net_backprop2.html ]
[ neural_net_backprop_batch_training.html ]
[ nn_single_lyr.html ]
[ sigmoid_calculator.html ]
[ svg_click_input_scaled_output.html ]
[ svg_input_for_neural_network_2e.html ]
[ two_lyr_neural_network.html ]

Just for fun, I created a Raspberry Pi 3B machine learning server. The Pi runs as a WIFI DHCP server and a VNC server. With this set up I connect to it with a tablet running a VNC client. The tablet only serves as a keyboard/mouse and monitor, all processes are running on the Pi. In this picture I am running Jupyter Notebook and Python 3. When I need the Pi to connect to the internet to download data, upload code, this and that, I run a small Bash script that renames my current configuration files. What mode I want to run the Pi depends on the parameter I pass to the Bash script.

Point of this server is that I can "play" even on a long flight or anywhere else where there is no internet connection.