The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
C4 Symmetry Pocket
C4 Symmetry Pocket tests architectural rotation invariance against two matched controls: canonical-only training and explicit rotation augmentation. Every variant uses the same base CNN and parameter count. The invariant network averages predictions across the four quarter-turn group elements.
The benchmark reports accuracy at 0, 90, 180, and 270 degrees, plus the worst angle. Group averaging requires four base-network evaluations at inference, so the robustness result is not presented as free computation.
Verified local result
All three variants contain exactly 1,201 parameters. Canonical-only training fell to 38.90% worst-angle accuracy, rotation augmentation reached 99.90%, and C4 group averaging reached 100% at every tested quarter turn. The invariant model requires four base-CNN evaluations per prediction.
uv run python projects/c4-symmetry-pocket/train.py
uv run pytest tests/test_c4_symmetry_pocket.py
- Downloads last month
- 3