demo.git - Vide-coded 64k demo system

Age	Commit message (Expand)	Author
18 hours	fix: Register cnn_conv1x1 snippet and add verification	skal
24 hours	fix: Move sigmoid activation to call site in CNN layer shader	skal
24 hours	fix: Replace clamp with sigmoid in CNN final layer	skal
25 hours	format .wgsl layer code (cosmetics)	skal
33 hours	opt: Move invariant in1 calculation outside CNN convolution loops	skal
35 hours	opt: Vec4-optimize CNN convolution shaders for SIMD	skal
35 hours	chore: Update CNN architecture to 3×3×3 with new trained weights	skal
37 hours	fix: Correct UV coordinate computation to match PyTorch linspace	skal
37 hours	fix: Add clamp to CNN final layer to match PyTorch training	skal
37 hours	refactor: Optimize CNN grayscale computation	skal
37 hours	update train_cnn.py and shader	skal
38 hours	fix: CNN training normalization pipeline consistency	skal
38 hours	udpate CNN shader code.	skal
39 hours	refactor: Optimize CNN normalization to eliminate redundant conversions	skal
40 hours	fix: Support variable kernel sizes in CNN layer generation	skal
41 hours	feat: CNN RGBD→grayscale with 7-channel augmented input	skal
42 hours	fix: Resolve CNN effect black screen bug (framebuffer capture + uniforms)	skal
45 hours	feat: Add multi-layer CNN support with framebuffer capture and blend control	skal
47 hours	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