docs(cnn_v3): full design doc — U-Net + FiLM architecture plan

- CNN_V3.md: complete design document
  - U-Net enc_channels=[4,8], ~5 KB f16 weights
  - FiLM conditioning (5D → γ/β per level, CPU-side MLP)
  - 20-channel feature buffer, 32 bytes/pixel: two rgba32uint textures
    - feat_tex0: albedo.rgb, normal.xy, depth, depth_grad.xy (f16)
    - feat_tex1: mat_id, prev.rgb, mip1.rgb, mip2.rgb, shadow, transp (u8)
  - 4-pass G-buffer: raster MRT + SDF compute + lighting + pack
  - Per-pixel parity framework: PyTorch / HTML WebGPU / C++ WebGPU (≤1/255)
  - Training pipelines: Blender full G-buffer + photo-only (channel dropout)
  - train_cnn_v3_full.sh spec (modelled on v2 script)
  - HTML tool adaptation plan from cnn_v2/tools/cnn_v2_test/index.html
  - Binary format v3 header spec
  - 8-phase ordered implementation checklist
- TODO.md: add CNN v3 U-Net+FiLM future task with phases
- cnn_v3/README.md: update status to design phase

handoff(Gemini): CNN v3 design complete. Phase 0 (stub G-buffer) unblocks
all other phases — one compute shader writing feat_tex0+feat_tex1 with
synthetic values from the current framebuffer. See cnn_v3/docs/CNN_V3.md
Implementation Checklist.

1 files changed, 17 insertions, 1 deletions

diff --git a/TODO.md b/TODO.md
index aca1257..0ced5e8 100644
--- a/TODO.md
+++ b/TODO.md
@@ -60,7 +60,23 @@ Ongoing shader code hygiene for granular, reusable snippets.
 
 ---
 
-## Future: CNN v3 8-bit Quantization
+## Future: CNN v3 — U-Net + FiLM
+
+U-Net architecture with FiLM conditioning. Runtime style control via beat/audio.
+Richer G-buffer input (normals, depth, material IDs). Per-pixel testability across
+PyTorch / HTML WebGPU / C++ WebGPU.
+
+**Prerequisites:** G-buffer implementation (GEOM_BUFFER.md)
+**Design:** `cnn_v3/docs/CNN_V3.md`
+
+**Phases:**
+1. G-buffer prerequisite
+2. Training infrastructure (Blender exporter + photo pipeline)
+3. WGSL shaders (enc/dec/bottleneck, deterministic ops)
+4. C++ effect class + FiLM uniform upload
+5. Parity validation (test vectors, ≤1/255 per pixel)
+
+## Future: CNN v2 8-bit Quantization
 
 Reduce weights from f16 (~3.2 KB) to i8 (~1.6 KB).