| Age | Commit message (Expand) | Author |
| 9 days | style: apply clang-formatHEADmain | skal |
| 9 days | fix: correct WGPUTexelCopyBufferInfo initialization order | skal |
| 9 days | fix: audio & effects cleanup — dead code removal, simplifications | skal |
| 9 days | fix: audio & effects cleanup — dead code removal, simplifications | skal |
| 9 days | fix: code review cleanup — bugs, dead code, factorization, simplification | skal |
| 9 days | fix: code review cleanup — bugs, dead code, factorization (-167 lines) | skal |
| 2026-05-14 | ans: order-0 rANS coder + WGSL asset compression | skal |
| 2026-03-29 | update weights | skal |
| 2026-03-29 | update assets and weights | skal |
| 2026-03-29 | more 3d assets for cnn_v3 | skal |
| 2026-03-29 | new script | skal |
| 2026-03-29 | docs: consolidate and sync docs with current codebase state | skal |
| 2026-03-29 | fix(cnn_v3): handle OPEN_EXR_MULTILAYER removal in Blender 5.x | skal |
| 2026-03-29 | docs(3d): document set_direct_render() in Y-flip rule section | skal |
| 2026-03-29 | fix(3d): restore correct orientation in test_3d_render (direct-to-surface) | skal |
| 2026-03-29 | feat(procedural): add plasma, voronoi, normalmap generators | skal |
| 2026-03-29 | fix: reorder WGPURequestAdapterOptions fields to match declaration order | skal |
| 2026-03-28 | fix(audio): fix early timing drift in tracker, use ola_decode_frame in synth | skal |
| 2026-03-27 | fix(cnn_v3): L1 loss + depth-grad tanh normalization to reduce flat convergence | skal |
| 2026-03-27 | fix(cnn_v3): remove dec0 ReLU, load FiLM MLP at runtime | skal |
| 2026-03-26 | fix(audio): P1-P3 fixes from audio code review | skal |
| 2026-03-26 | update the weights | skal |
| 2026-03-26 | chore(scripts): document and guard mingw-w64 setup | skal |
| 2026-03-26 | fix(src/platform): code review cleanup | skal |
| 2026-03-26 | chore(src/gpu): remove stale cnn_v1/v2 artifacts and dead comments | skal |
| 2026-03-26 | fix(cnn_v3/training): fix defaults and help strings across py tools | skal |
| 2026-03-26 | feat(cnn_v3): upgrade architecture to enc_channels=[8,16] | skal |
| 2026-03-25 | fix(cnn_v3/training): rebuild optimizer before loading state on resume past F... | skal |
| 2026-03-25 | fix(cnn_v3/tools): rename Output→Dec0 in viz panel; fix BN weight cnt 72→584 | skal |
| 2026-03-25 | feat(cnn_v3/training): load prev.png when available; document web tool prev gap | skal |
| 2026-03-25 | feat(cnn_v3): 3×3 dilated bottleneck + Sobel loss + FiLM warmup + architectu... | skal |
| 2026-03-25 | update weights | skal |
| 2026-03-25 | feat(cnn_v3/training): add --single-sample option + doc fixes | skal |
| 2026-03-25 | feat(cnn_v3): add infer_cnn_v3.py + rewrite cnn_test for v3 parity | skal |
| 2026-03-25 | feat(cnn_v3/tools): embed default weights in HTML tool; add --html export flag | skal |
| 2026-03-25 | update weights | skal |
| 2026-03-24 | add more samples + tool html | skal |
| 2026-03-24 | move and improve the 'adjust.html' tool | skal |
| 2026-03-24 | fix(fft): make bit_reverse_permute static, remove from public API and tests | skal |
| 2026-03-24 | fix(fft): replace iterative twiddle with direct cosf/sinf, add tests A-E | skal |
| 2026-03-23 | test(fft): re-enable DCT tests, document twiddle accumulation bug | skal |
| 2026-03-23 | add a registration tool | skal |
| 2026-03-23 | docs: update temporal feedback docs — wire_dag auto-wiring, F16X8 format | skal |
| 2026-03-23 | fix(cnn_v3_debug): add CNNv3Effect to debug sequence for prev.r/g/b temporal ... | skal |
| 2026-03-23 | feat(gbuffer): wire_dag() + find_downstream_output() for temporal feedback | skal |
| 2026-03-23 | feat(cnn_v3): GBufferEffect temporal feedback via post_render() | skal |
| 2026-03-23 | feat(cnn_v3): shadow→dif migration complete (ch18) | skal |
| 2026-03-23 | wip(cnn_v3): shadow→dif intermediate + scene tweaks + migration plan | skal |
| 2026-03-22 | refactor(cnn_v3): simplify sphere SDF in shadow pass, remove per-frame alloc | skal |
| 2026-03-22 | fix(cnn_v3): shadow pass — 5 bugs fixed, labels in gbuf_view | skal |