Basic playground exercises that I used to learn how to do image processing in WebGPU, using help from Google Code Labs, and WebGPU Fundamentals
The repo is divided into a few basic test exercises in this order
- 1 - Render a black screen to a canvas (Hello world)
- 2 - Render a red square on top of a black screen
- texture-loading - Load an image texture into the canvas and display it
- yuv - Render a convert an image texture from RGB colorspace to YUV
- gaussian - Apply a gaussian blur to an image texture
- sequence - Apply a sequence of image operations (convert to YUV in the first render pass, then apply gaussian blur in the second render pass) and paint the result to screen
Working up to sequence executution of various image processing tasks should be enough of a basis to begin actully writing a basic neural network implementation in WebGPU (Coming soon, in a seperate repository)