autograd: Add VJP and JVP rules for aten::aminmax#158241

Open

vijayabhaskar-ev wants to merge 12 commits intopytorch:mainfrom

vijayabhaskar-ev:fix-aminmax-autograd-rules

Commits on Jul 15, 2025

autograd: Add VJP and JVP rules for aten::aminmax
vijayabhaskar-ev
committed
- Updated autograd rules in FunctionsManual.cpp and FunctionsManual.h to handle amin and amax correctly.
vijayabhaskar-ev
committed
[autograd] Optimize tensor allocation in aminmax_backward
vijayabhaskar-ev
committed
• Modified the mask generation logic.
vijayabhaskar-ev
committed
Refactored aminmax_backward and added testcases for amin, amax and aminmax
vijayabhaskar-ev
committed
Update dimension parameter to use std::optional<int64_t> and refactored testcase
vijayabhaskar-ev
committed
Merge branch 'fix-aminmax-autograd-rules' of https://github.com/vijayabhaskar-ev/pytorch into fix-aminmax-autograd-rules
vijayabhaskar-ev
committed
Added restore_reduced_dims for grad_max and grad_min
vijayabhaskar-ev
committed

Commits on Jul 16, 2025

Removed restore_reduced_dims fro grad_min and grad_max in aminmax_backward
vijayabhaskar-ev
committed

Commits on Jul 17, 2025

Refactored aminmax_backward and added reduced_dim for grad_min and grad_max
vijayabhaskar-ev
committed

Commits on Jul 21, 2025