diff options
Diffstat (limited to 'TODO.md')
| -rw-r--r-- | TODO.md | 17 |
1 files changed, 10 insertions, 7 deletions
@@ -24,22 +24,25 @@ Self-contained workspaces for parallel demo development. --- -## Priority 2: CNN v2 - Parametric Static Features (Task #85) [PLANNING] +## Priority 2: CNN v2 - Parametric Static Features (Task #85) [IN PROGRESS] Enhanced CNN post-processing with multi-dimensional feature inputs. **Design:** `doc/CNN_V2.md` -**Implementation phases:** -1. Static features compute shader (RGBD + UV + sin encoding + bias) -2. C++ effect class (multi-pass layer execution) -3. Training pipeline (PyTorch f32 → f16 export) -4. Validation tooling (end-to-end checkpoint testing) +**Status:** +- ✅ Phase 1: Static features shader (RGBD + UV + sin encoding + bias → 8×f16, 3 mip levels) +- ✅ Phase 2: C++ effect class (CNNv2Effect skeleton, multi-pass architecture) +- ✅ Phase 3: Training pipeline (`train_cnn_v2.py`, `export_cnn_v2_shader.py`) +- ✅ Phase 4: Validation tooling (`scripts/validate_cnn_v2.sh`) +- ⏳ Phase 5: Full implementation (bind groups, multi-pass execution, layer shaders) + +**Next:** Complete CNNv2Effect render pipeline, test with trained checkpoint **Key improvements over v1:** - 7D static feature input (vs 4D RGB) - Per-layer configurable kernels (1×1, 3×3, 5×5) -- Float16 weight storage (~6.4 KB vs 3.2 KB) +- Float16 weight storage (~6.4 KB) **Target:** <10 KB for 64k demo constraint |