summaryrefslogtreecommitdiff
AgeCommit message (Expand)Author
21 hoursfix(timeline-editor): playback indicator zoom tracking and UX improvementsskal
21 hoursfeat(timeline-editor): consolidate playback indicator positioning and improve UXskal
22 hoursfix(shader): add dfWithID to sdf_test.wgsl for test compatibilityskal
22 hoursfix(cli): enable WAV dump without DEMO_HEADLESS requirementskal
22 hoursfix(cli): change --dump_wav to --dump-wav and add error handlingskal
23 hoursfeat(gpu): add shader labels for better error identificationskal
23 hoursfeat(spectral-editor): add waveform intensity viewer for sample offsetskal
37 hoursfeat(gpu): add two-pass raymarching infrastructureskal
37 hoursfeat(gpu): add SDF camera infrastructure and effect base classskal
38 hoursdocs: add audio system enhancement tasksskal
38 hoursdocs: mark tracker humanization task as IMPLEMENTEDskal
38 hoursfeat(tracker): add sample offset and humanizationskal
38 hourssome tweakingskal
40 hoursrefactor(wgsl): consolidate SDF shapes into single common fileskal
41 hoursrefactor(wgsl): replace inline fullscreen_vs with common includeskal
41 hoursstyle: Apply clang-format to codebaseskal
41 hoursrefactor(wgsl): modularize common shader functionsskal
41 hoursrefactor(gpu): Relocate effects to src/effects and streamline includesskal
41 hoursfeat: Move test_platform.cc to util subdirectoryskal
41 hoursdocs: Update agent state from project documentationskal
42 hoursRefactor: extract setup_standard_textures helperskal
42 hoursRefactor: add gpu_create_post_process_texture helperskal
42 hoursRefactor: add gpu_create_texture_view_2d helperskal
43 hoursDocs: WGPU helper functions referenceskal
43 hoursRefactor: factorize WGPU init code for 3D testsskal
43 hoursFix: CNN v2 compute shader validation errorskal
43 hoursRefactor: factorize common WGPU patterns into helper functionsskal
2 daysCNN v2: bilinear mip-level sampling and UI improvementsskal
2 daysFactor common test patterns into reusable utilitiesskal
2 daysRemove non-functional Gantt test scriptsskal
2 daysUpdate docs: CNN v2 sigmoid activation summaryskal
2 daysReplace hard clamp with sigmoid activation in CNN v2skal
2 daysAdd --lr parameter to CNN v2 training pipelineskal
2 daysDocument CNN v2 training pipeline improvementsskal
2 daysStreamline CNN v2 training pipeline outputskal
2 daysFix CNN v2 training: always save final checkpoint, derive num_layersskal
2 daysAdd --output-weights option to CNN v2 training pipelineskal
2 daysFix --mix option: blend prev layer with static p4-p7, not p0-p3skal
2 daysFix CNN v2 static feature channel mapping (p4-p7 → channels 8-11)skal
2 daysCNN v2 web tool: Remove vizScale, always show raw layer valuesskal
2 daysCNN v2 web tool: Fix vizScale for final layer previewskal
2 daysgen_identity_weights: Change --mix to 50-50 blendskal
2 daysgen_identity_weights: Add --p47 option for static feature visualizationskal
2 daysgen_identity_weights: Add --mix option for static feature blendingskal
2 daysCNN v2 tool: Fix off-by-one in composited layer filenamesskal
2 daysCNN v2: Fix weight buffer offset bugskal
2 dayscnn_test: --weights now overrides layer config from .bin fileskal
2 dayscnn_test: Add --weights option for runtime weight loadingskal
2 daysCNN v2: Remove vizScale, always clip to [0,1]skal
2 daysCNN v2: Fix Layer 0 visualization scale (was 0.5, now 1.0)skal