demo.git - Vide-coded 64k demo system

Age	Commit message (Expand)	Author
22 hours	fix: Move sigmoid activation to call site in CNN layer shader	skal
22 hours	fix: Replace clamp with sigmoid in CNN final layer	skal
23 hours	feat: Add early stopping to CNN training	skal
23 hours	fix: CNN training/inference to match WGSL sliding window	skal
23 hours	format .wgsl layer code (cosmetics)	skal
24 hours	fix: Guard cnn_test build with STRIP_ALL check	skal
24 hours	refactor: Modularize CMake build system into 10 specialized modules	skal
24 hours	fix: CNN test tool GPU readback with wgpuDevicePoll	skal
25 hours	debug: Add shader load verification and GPU sync to CNN test tool	skal
25 hours	fix: CNN test tool ping-pong bug and RGBA16Float intermediates	skal
25 hours	feat: Add CNN shader testing tool with GPU texture readback	skal
32 hours	fix: Use patch-based inference to match CNN training distribution	skal
32 hours	opt: Move invariant in1 calculation outside CNN convolution loops	skal
33 hours	opt: Vec4-optimize CNN convolution shaders for SIMD	skal
33 hours	chore: Update CNN architecture to 3×3×3 with new trained weights	skal
33 hours	docs: Update CNN training documentation with patch extraction	skal
33 hours	feat: Add salient-point patch extraction for CNN training	skal
35 hours	fix: Use ClampToEdge sampler for CNN to avoid edge wrapping	skal
35 hours	fix: Correct UV coordinate computation to match PyTorch linspace	skal
35 hours	fix: Add clamp to CNN final layer to match PyTorch training	skal
35 hours	fix: PostProcessHelperTest signature and fixture sharing	skal
35 hours	refactor: Optimize CNN grayscale computation	skal
35 hours	update train_cnn.py and shader	skal
36 hours	docs: Add inference mode to training documentation	skal
36 hours	feat: Add inference mode to train_cnn.py for ground truth generation	skal
36 hours	fix: CNN training normalization pipeline consistency	skal
37 hours	udpate CNN shader code.	skal
37 hours	feat: Add --shader-only option to convert_shadertoy.py	skal
37 hours	refactor: Optimize CNN normalization to eliminate redundant conversions	skal
38 hours	update timeline.seq	skal
38 hours	fix: Flip Y-axis to match ShaderToy coordinate convention	skal
38 hours	refactor: Improve convert_shadertoy.py to generate compile-ready code	skal
38 hours	feat: Add Scene1 effect from ShaderToy (raymarching cube & sphere)	skal
38 hours	chore: Remove incomplete CubeSphere effect	skal
38 hours	docs: Fix EFFECT keyword syntax and add automation-friendly workflow	skal
38 hours	feat: Add ShaderToy conversion tools	skal
39 hours	fix: Support variable kernel sizes in CNN layer generation	skal
39 hours	docs: Document WGPU builder refactoring in COMPLETED.md	skal
39 hours	refactor: Factor WGPU boilerplate into builder pattern helpers	skal
39 hours	feat: CNN RGBD→grayscale with 7-channel augmented input	skal
40 hours	udpate	skal
40 hours	update timeline	skal
40 hours	fix: Resolve CNN effect black screen bug (framebuffer capture + uniforms)	skal
42 hours	fix: Capture scene framebuffer before post-processing for CNN effect	skal
43 hours	feat: Add multi-layer CNN support with framebuffer capture and blend control	skal
46 hours	docs: Update and streamline CNN training documentation	skal
46 hours	feat: Add checkpointing support to CNN training script	skal
46 hours	fix: Auto-expand single kernel size to all layers in training script	skal
46 hours	update target images	skal
46 hours	feat: Add coordinate-aware CNN layer 0 for position-dependent stylization	skal