demo.git - Vide-coded 64k demo system

Age	Commit message (Expand)	Author
33 hours	fix: Move sigmoid activation to call site in CNN layer shader	skal
33 hours	fix: Replace clamp with sigmoid in CNN final layer	skal
33 hours	format .wgsl layer code (cosmetics)	skal
42 hours	opt: Move invariant in1 calculation outside CNN convolution loops	skal
43 hours	opt: Vec4-optimize CNN convolution shaders for SIMD	skal
44 hours	chore: Update CNN architecture to 3×3×3 with new trained weights	skal
45 hours	fix: Correct UV coordinate computation to match PyTorch linspace	skal
45 hours	fix: Add clamp to CNN final layer to match PyTorch training	skal
46 hours	refactor: Optimize CNN grayscale computation	skal
46 hours	update train_cnn.py and shader	skal
47 hours	fix: CNN training normalization pipeline consistency	skal
47 hours	udpate CNN shader code.	skal
48 hours	refactor: Optimize CNN normalization to eliminate redundant conversions	skal
2 days	fix: Support variable kernel sizes in CNN layer generation	skal
2 days	feat: CNN RGBD→grayscale with 7-channel augmented input	skal
2 days	fix: Resolve CNN effect black screen bug (framebuffer capture + uniforms)	skal
2 days	feat: Add multi-layer CNN support with framebuffer capture and blend control	skal
2 days	feat: Add coordinate-aware CNN layer 0 for position-dependent stylization	skal
2 days	feat: Add CNN post-processing effect with modular WGSL architecture	skal