MLBlocks provides a visual, notebook-like block canvas for assembling Python ML workflows. Users can drag predefined blocks, edit inline or in a full editor, rewire the control-flow chain, and export to a runnable main.py.
Run Program
Navigate inside a folder
Press f5
Click ctrl + shift + p
Type MLBlocks: Open Canvas
Wait 5 seconds
Click ctrl + shift + p
Type MLBlocks: Open Canvas
More Commands
MLBlocks: Open Canvas
MLBlocks: Convert Project
MLBlocks: Run Project
MLBlocks: Debug Project
MLBlocks: Stop Execution
Features
This extension provides an easy digestable way for handling large code repositories, as well as aiding in ML Development
Right-Click selct "Custom Block" to create a fully custom block (Rename Header/Output type)
Use the Left pane to drag and drop blocks of your choice
Connect blocks using the wire tool (Output (Red) --> Input (Green))
The Execution Order tab will show the flow of the program
There are two ways to run the program (similiar to a .py file and a .ipynb file)
To replicate a .py file execution
Connect all blocks together (Every block must be connected) [If the execution order does not show up, you have an invalid/missing connection]
You can click the "Open Editor" button on a block, to open a seperate file to edit your block in a larger view [ctrl+s to save, change will be applied in real-time]
Click the "Convert" button at the top, which will translate the blocks to .py files in the "ml_blocks_output"
Click the "Run" button to see execution
To replicate a .ipynb file execution
No need to connect all blocks
You can select the "Run Block" to just run that piece of code
You can link several blocks together, click "Run Block" on a downstream block, and all previous blocks up to (including) the current block will run [In events a block relies on a variable or structure of a previous block]
