Build Scratch 3.0 extensions for adding tensorflow.js function

Build a Scratch 3.0 extension for adding [login to view URL] function for kids learning about the [login to view URL] (TFJS). A. Build a scratch 3 extension for TFJS - The general direction: - Building a scratch 3.0 extension for [login to view URL] ([login to view URL]) like in kittenblock ([login to view URL]) - So that it is a number of blocky functions to make use of the tensorflow js function - Load CSV [in csv] to [] - Load CSV [in csv] to [] col [0,1,2,3] - Load Image [] xs [] label [] - Shuffle xs [] ys [] - Save Model [] name [mnist] - Load Model [in json] Weight [in file extension bin] - Init model input shape [] - Model [] add dense unit [] act [] - Model [] add Conv2D Kernel [] filters [] act [] - Model [] add MaxPool 2D she [] - Model [] add Flatten Layer - Model [] compile opt [] loss [] learn rate [] - Model [] Fit [] epochs - When Train finish - Model [] Predict [] - Model [] Predict [] by Label [] - Set xs [] - Set xs [] shape [] - Image xs [] channel [] size [] shape [] - Set output[] - Set output Label [] - Set Output Label [] size [] - Pen Graph - Stage Graph - Video Graph - Show Debug Panel - Debug Image [] - Debug off When the user press “`” button, a debug mode for machine learning model is coming out. - The extension can at least build following: - Linear Regression - Image Classification - MNIST Hand-written recognition - Save & load MNIST model - Mobilenet for object identification - Mobilenet w/ robotic B. teach me and make a guide for hosting the scratch 3.0 in cloud AWS and Digital Ocean with load balance: - Grab the necessary repositories for scratch 3.0 from [login to view URL] (At least you need the scratch-gui) - The general direction: - 1) Running it in local - Refer to the [login to view URL] from scratch-gui - npm start - localhost:8601 - 2) Compile and build the GUI items - Using web pack to compile the GUI items. - npm run build - Get the whole project from “build” folder. - 3) Host it into a cloud ==> That’s I need you to do. C. Build a login in system like in [login to view URL] - Profile page as in the [login to view URL] - The user can store, load and delete the project they have done under his/her profile - Account setting for Password, account name, email