how to get memory allocate detail in model infer?

[zhangvia]

Describe the issue

i want to analysis the memory cost in model infer, but when i set enable_profiling=true, there is nothing about memory detail in profile json. do i need build onnxruntime from source with onnxruntime_ENABLE_MEMORY_PROFILE?

To reproduce

import onnxruntime as ort
model_path = "/media/general.onnx"
sess_options = ort.SessionOptions()
sess_options.enable_profiling = True
model_sess = ort.InferenceSession(model_path,sess_options,providers=[('CUDAExecutionProvider',{'device_id':1})])
import numpy as np
import numpy as np
input = {"input":np.random.rand(1,3,1170,2532).astype(np.float32),"orig_im_size":[1170,2532]}
model_sess.run(None,input)

Urgency

No response

Platform

Linux

OS Version

ubuntu 20.04

ONNX Runtime Installation

Released Package

ONNX Runtime Version or Commit ID

1.18.1

ONNX Runtime API

Python

Architecture

X64

Execution Provider

CUDA

Execution Provider Library Version

cuda 12.4

[tianleiwu]

Use --enable_memory_profile in build.
See #5658

[zhangvia]

Use --enable_memory_profile in build. See #5658

i build onnxruntime from source with args ./build.sh --use_cuda --cudnn_home /usr/local/cuda --cuda_home /usr/local/cuda --enable_nvtx_profile --enable_memory_profile --allow_running_as_root --build_wheel --cmake_extra_defines CMAKE_POLICY_VERSION_MINIMUM=3.5

and there is no error in build process, but seg fault happened in some pytorch test. and there is no whl file in build directory. how can i install python interface of onnxruntime after build?