SOLVED! I was using Ints to access the Image, it should have been floats! Changes to the pixel shader: I added "cast to float" // uniform vec2 size; // #define BUFFER_SIZE 100 uniform int height = 1; uniform int width = 1; uniform Image bias; // x output uniform Image intake; // y input un...

In this case, the Textures are in pixels(int) accessed. I changed the pixel shader uniform int height = 1; uniform int width = 1; uniform Image bias; // x output uniform Image intake; // y input uniform Image weights; // x is output y is input // x + M * y float activation(float value){ return 1 / (...

I know, that LÖVE 12 has compute shaders. But my questioning was about, if there is a workaround to my problem in the current version of LÖVE. The GPU is receiving "0 | 0.8 | 0.8 | 0.8" as an Image, but the Image that was send has data of "0 | 0.2 | 0.4 | 0.8" (both times referri...

I was trying to make an compute shader using the pixel shader, but the Image is not correctly sendet to the GPU. Using LÖVE 11.4. main.lua cmp = love.graphics.newShader("compute.cmp") test = false MAX = 10 SIZE = 16 function sendData(input,weight,bias,output) -- float[] -- input = Image[y ...