Describe the bug
ttnn.Maxpool2d fails to produce result for some test cases while working on Blazepose model.
To Reproduce
Steps to reproduce the behavior:
pick the test from the commit
Run pytest <path to file>
Expected behavior
As the above shared configuration works in pytorch, it is expected to run in ttnn aswell.
Error
E RuntimeError: TT_FATAL @ ../ttnn/cpp/ttnn/operations/pool/maxpool/device/max_pool_program_factory.cpp:30: is_pow2
E info:
E Row size (nchannels * bytes = 192) should be power of 2 (false).
E backtrace:
E --- tt::tt_metal::MaxPool::validate(std::__1::vector<tt::tt_metal::Tensor, std::__1::allocator<tt::tt_metal::Tensor>> const&) const
Describe the bug ttnn.Maxpool2d fails to produce result for some test cases while working on Blazepose model.
To Reproduce Steps to reproduce the behavior:
pick the test from the commit Run
pytest <path to file>
Expected behavior As the above shared configuration works in pytorch, it is expected to run in ttnn aswell.
Error