mirror of
https://github.com/MLSysBook/TinyTorch.git
synced 2026-03-09 15:02:25 -05:00
Fixes after merge conflicts: - Fix tensor reshape error message format - Fix __init__.py imports (remove BatchNorm2d, fix enable_autograd call) - Fix attention mask broadcasting for multi-head attention - Fix memoization module to use matmul instead of @ operator - Fix capstone module count_parameters and CosineSchedule usage - Add missing imports to benchmark.py (dataclass, Profiler, platform, os) - Simplify capstone pipeline test to avoid data shape mismatch All 20 modules now pass tito test --all