When I ran the tvm_pytorch_resnet18_inference.py, I met the problem about tvm_mlir_learn HOT 5 CLOSED

I met the same problem, how did you solve it?
thanks

from tvm_mlir_learn.

/tvm/build/config.cmake
set(USE_LLVM "/usr/bin/llvm-config --link-static")
set(USE_MICRO OFF)
/tvm/CMakelist.txt
set(HIDE_PRIVATE_SYMBOLS ON)\
cmake -DCMAKE_BUILD_TYPE=Debug ../
make -j16
I modify the above .

from tvm_mlir_learn.

these modifications worked
I have a new puzzle.
In terms of reasoning time, relay reasoning time is longer, which does not reflect the optimization effect of TVM?

from tvm_mlir_learn.

these modifications worked I have a new puzzle. In terms of reasoning time, relay reasoning time is longer, which does not reflect the optimization effect of TVM?

same question.
Relay time: 0.37134288150118666 seconds
Torch time: 0.0024548148998292164 seconds

from tvm_mlir_learn.

these modifications worked I have a new puzzle. In terms of reasoning time, relay reasoning time is longer, which does not reflect the optimization effect of TVM?

You should use AutoTVM or Ansor to optimize it. Please refer to : https://tvm.apache.org/docs/tutorial/autotvm_relay_x86.html & https://tvm.apache.org/docs/tutorial/auto_scheduler_matmul_x86.html

from tvm_mlir_learn.