Hello! this is my nano-cat 1 model! it takes image data from the project https://scratch.mit.edu/projects/1369041956 to then use the model to use very simple diffusion. instead of the ai viewing its whole image, I coded it to have the training data's pixel values and nano goes through all pixels of his random noise and slightly modifies it until it is correct. then it is displayed. im wanting to fnd a way to make another model to understand the prompt and generate the image data, while the nano-cat draws it from the data (hopefully soon nano will have creativity to draw more)
thanks to @Mryellowdoggy for the image rendering engine thanks to Google's nano banana for the idea thanks to Google's nano banana for the ai math thanks to opera for making this easy!
