8000 minipytorch/TODO.md at main · amartya116/minipytorch · GitHub
[go: up one dir, main page]

Skip to content

Latest commit

 

History

History
14 lines (11 loc) · 876 Bytes

TODO.md

File metadata and controls

14 lines (11 loc) · 876 Bytes
  • Confirm current build/run target(s) for linear.cpp (linear target vs minipytorch.exe)

  • Fix header include errors by removing .cpp includes from:

    • src/cpp/autograd/autogradnodes/addnode.hpp
    • src/cpp/autograd/autogradnodes/mulnode.hpp
    • src/cpp/autograd/autogradnodes/activationnode.hpp
    • src/cpp/autograd/autogradnodes/lossfunctionnode.hpp
  • Ensure headers include correct forward declarations / includes for TensorImpl and Dtypes.

  • Rebuild with cmake --build build --config Release --target linear and verify compilation (CMake target not available; MSBuild expects other targets instead)

  • Run the produced executable and verify printed output matches expected in linear.cpp

  • If output mismatch: inspect autograd backward for MSE + matmul/add gradients and optimizer update logic

  • Update TODO progress accordingly

0