Error in installing TVM on zcu104

Is there a new installation process?

After installing pyxi on Xilinx zcu104 FPGA, error “could not load CPU TF runtime” was reported when installing TVM. The specific error is as follows:

xilinx@pynq:~/tvm/python$ cd …/build/

xilinx@pynq:~/tvm/build$ cmake …

– Build with RPC support…

– Build with Graph Executor support…

– Build with profiler…

– VTA build with VTA_HW_PATH=/home/xilinx/tvm/3rdparty/vta-hw

– Build VTA runtime with target: sim

– Build with contrib.random

– Build with contrib.sort

– Build with contrib.hybriddump

/usr/local/lib/python3.6/dist-packages/pyxir-0.1.6-py3.6-linux-aarch64.egg/pyxir/runtime/init.py:34: UserWarning: Could not load cpu-tf runtime because of error: No module named 'tensorflow’ ** .format(e))**

– Build with contrib.vitisai

– Git found: /usr/bin/git

– Found TVM_GIT_COMMIT_HASH=1fb32b055171f7b82b506b0f8e3096db1acc180c – Autoset: USE_LIBBACKTRACE=ON in Linux

– Building with libbacktrace…

– Building with TVM Map…

– Build with thread support…

– Configuring done

– Generating done

– Build files have been written to: /home/xilinx/tvm/build

@williamyang4978

You will see a warning about the ‘cpu-tf’ runtime not being found. This warning is expected on the board and can be ignored. Note also that you shouldn’t import the PyXIR targets in the run script (import pyxir.contrib.target.DPUCZDX8G ).

You can find the same info from this page.