demo.git/cnn_v3/src, branch main

style: apply clang-format

2026-05-21T06:10:47Z

fix: code review cleanup — bugs, dead code, factorization, simplification

2026-05-20T21:21:59Z

Bugs: - B1: fix dead tempo debug (prev_tempo captured after assignment) - B2: fix ReloadAssetsFromFile leak for disk-loaded assets; simplify DropAsset - B3: fix get_free_pool_slot leak (unregister synth + free data on reuse) - B4: volatile -> std::atomic with acquire/release in miniaudio_backend, synth - B5: fix unaligned reads in scene_loader (memcpy-based read_f32/read_u32) - B6: fix shader module + BGL + pipeline layout leaks in gpu.cc, pipeline_builder Dead code: - D1: remove unused particle_defs.h - D3: remove create_post_process_pipeline_simple (zero callers) - D4: remove empty gpu_draw() - D5: remove write-only Hybrid3D::initialized_ - D6: remove legacy pending buffer path in audio.cc Factorization: - F1: Effect::run_fullscreen_pass() replaces boilerplate in 5 effects - F2: particle_common.wgsl snippet, #include in 3 WGSL shaders - F3: gpu_create_shader_module() helper, used in 3 call sites - F5: get_world_aabb() shared between bvh.cc and physics.cc - F6: samples_to_seconds() replaces 6 inline expressions - F7: gpu_create_linear/nearest_sampler use SamplerCache; add nearest() preset Simplification: - S9+S1: WgslSamplerType param; Scene2Effect collapsed to thin wrapper - S4: FFT heap allocs -> stack arrays (zero allocs on hot path) - S5: ObjectType::CUBE documented as legacy alias for BOX; default changed - S6: bind group dirty-flag in Renderer3D; remove duplicate pipeline set - S7: create_gpu_procedural() helper in texture_manager (~80 lines removed) 37/37 tests passing. handoff(Claude): code review batch — all items verified, no regressions.

fix(cnn_v3): remove dec0 ReLU, load FiLM MLP at runtime

2026-03-27T06:59:00Z

Two bugs blocking training convergence: 1. dec0 ReLU before sigmoid constrained output to [0.5,1.0] — network could never produce dark pixels. Removed F.relu in train_cnn_v3.py and max(0,…) in cnn_v3_dec0.wgsl. Test vectors regenerated. 2. set_film_params() used hardcoded heuristics instead of the trained MLP. Added CNNv3FilmMlp struct + load_film_mlp() to cnn_v3_effect.h/.cc. MLP auto-loaded from ASSET_WEIGHTS_CNN_V3_FILM_MLP at construction; Linear(5→16)→ReLU→Linear(16→72) runs CPU-side each frame. 36/36 tests pass. Parity max_err=4.88e-4 unchanged. handoff(Gemini): retrain from scratch — needs ≥50 samples (currently 11). See cnn_v3/docs/HOWTO.md §2-3.

feat(cnn_v3): upgrade architecture to enc_channels=[8,16]

2026-03-26T06:03:01Z

Double encoder capacity: enc0 4→8ch, enc1 8→16ch, bottleneck 16→16ch, dec1 32→8ch, dec0 16→4ch. Total weights 2476→7828 f16 (~15.3 KB). FiLM MLP output 40→72 params (L1: 16×40→16×72). 16-ch textures split into _lo/_hi rgba32uint pairs (enc1, bottleneck). enc0 and dec1 textures changed from rgba16float to rgba32uint (8ch). GBUF_RGBA32UINT node gains CopySrc for parity test readback. - WGSL shaders: all 5 passes rewritten for new channel counts - C++ CNNv3Effect: new weight offsets/sizes, 8ch uniform structs - Web tool (shaders.js + tester.js): matching texture formats and bindings - Parity test: readback_rgba32uint_8ch helper, updated vector counts - Training scripts: default enc_channels=[8,16], updated docstrings - Docs + architecture PNG regenerated handoff(Gemini): CNN v3 [8,16] upgrade complete. All code, tests, web tool, training scripts, and docs updated. Next: run training pass.

feat(cnn_v3): 3×3 dilated bottleneck + Sobel loss + FiLM warmup + architecture PNG

2026-03-25T09:05:42Z

- Replace 1×1 pointwise bottleneck with Conv(8→8, 3×3, dilation=2): effective RF grows from ~13px to ~29px at ¼res (~+1 KB weights) - Add Sobel edge loss in training (--edge-loss-weight, default 0.1) - Add FiLM 2-phase training: freeze MLP for warmup epochs then unfreeze at lr×0.1 (--film-warmup-epochs, default 50) - Update weight layout: BN 72→584 f16, total 1964→2476 f16 (4952 B) - Cascade offsets in C++ effect, JS tool, export/gen_test_vectors scripts - Regenerate test_vectors.h (1238 u32); parity max_err=9.77e-04 - Generate dark-theme U-Net+FiLM architecture PNG (gen_architecture_png.py) - Replace ASCII art in CNN_V3.md and HOW_TO_CNN.md with PNG embed handoff(Gemini): bottleneck dilation + Sobel loss + FiLM warmup landed. Next: run first real training pass (see cnn_v3/docs/HOWTO.md §3).

fix(cnn_v3_debug): add CNNv3Effect to debug sequence for prev.r/g/b temporal feedback

2026-03-23T07:05:12Z

timeline.seq is the canonical source — timeline.cc was wrongly hand-edited. Add CNNv3Effect + cnn_out (gbuf_albedo) node to cnn_v3_debug sequence so wire_dag() can wire GBufferEffect.cnn_output_node_ correctly. Also fix node_prev_tex_ NodeType: F16X8 (Rgba16Float+CopyDst) to match CNNv3Effect output format (GBUF_ALBEDO = Rgba16Float). Regenerated timeline.cc via: python3 tools/seq_compiler.py workspaces/main/timeline.seq Co-Authored-By: Claude Sonnet 4.6

feat(gbuffer): wire_dag() + find_downstream_output() for temporal feedback

2026-03-23T06:54:18Z

- Add Effect::wire_dag() virtual (called from init_effect_nodes after full DAG built) - Add Effect::find_downstream_output() protected helper (first downstream consumer output) - GBufferEffect::wire_dag() auto-sets cnn_output_node_ via find_downstream_output, guarding against sink (external view, null texture) - GBufferEffect::post_render() null-checks src texture before CopyTextureToTexture - Tests: find_downstream_output cases + wire_dag integration in test_effect_base - Doc: SEQUENCE.md updated with wire_dag pattern, helper contract, and sink guard Co-Authored-By: Claude Sonnet 4.6

feat(cnn_v3): GBufferEffect temporal feedback via post_render()

2026-03-23T06:31:14Z

- Add Effect::post_render() virtual hook, called after all effects in the sequence have rendered each frame. Default is no-op. - Sequence::render_effects() runs a second pass invoking post_render() on all DAG nodes after the render pass completes. - GBufferEffect: declare internal node_prev_tex_ (U8X4_NORM) for persistent prev-frame CNN output. post_render() copies cnn_output_node_ → node_prev_tex_ via CopyTextureToTexture. render() binds node_prev_tex_ as prev_cnn (binding 6) — zero on frame 0 (matches training convention). - Expose set_cnn_output_node(name) API; call once at setup. - Drop brittle ping-pong / input_nodes_[0] fallback. - Update doc/SEQUENCE.md: post_render() semantics, frame execution order, temporal feedback canonical pattern, node types table with G-buffer types. - Update cnn_v3/docs/HOWTO.md: temporal feedback wiring section. 36/36 tests passing. handoff(Gemini): prev.rgb temporal feedback now correct and generic. Set set_cnn_output_node("sink") (or CNN output node name) once at setup.

wip(cnn_v3): shadow→dif intermediate + scene tweaks + migration plan

2026-03-22T23:26:52Z

- gbuf_shadow.wgsl: normal bias 0.05→0.02 - gbuf_pack.wgsl: compute dif=diffuse*shadow, drop shadow from t1.z, store dif in t1.w (INTERMEDIATE — incorrect packing, see migration plan) - gbuf_deferred.wgsl: read dif from t1.w.x (matches intermediate packing) - gbuf_view.wgsl: expand to 4×6 grid, show dif.r/g/b in row 5 (INTERMEDIATE — to be reverted to 4×5 with ch18=dif) - gbuffer_effect.cc: add small hovering sphere (r=0.6) above scene; swap cube/sphere positions; both spheres pulsate - docs/GBUF_DIF_MIGRATION.md: full migration plan with checklist handoff(Claude): intermediate commit — GBUF_DIF_MIGRATION.md §Current State describes what is wrong and the full implementation checklist (5 steps).

refactor(cnn_v3): simplify sphere SDF in shadow pass, remove per-frame alloc

2026-03-22T22:51:40Z

gbuf_shadow.wgsl — dfWithID(): - Sphere: replace inv_model local-space transform with direct world-space formula (length(p - center) - radius). Exact, no matrix multiply, no floating-point error from matrix inversion that can corrupt soft-shadow penumbra over 64 march steps. - lp/scale now computed only inside the cases that need them (box/torus/plane) instead of eagerly for every object. gbuffer_effect.cc — upload_scene_data(): - Replace per-frame std::vector heap allocation with a file-static staging buffer s_obj_staging[256]: zero alloc per frame. handoff(Gemini): sphere SDF now exact; shadow march should be cleaner.