Question about Chakra Trace Generated by PyTorch

Hello, I am a beginner just starting to learn. I really appreciate the great work your team does with Chakra.

I am following the tutorial: https://github.com/mlcommons/chakra/wiki/Running-Simulation-with-Chakra
Regarding "Using Chakra Execution Trace and Kineto Traces Generated By PyTorch," I encounter an error whenever I reach the section in the purple box:

The error is shown below:

(astra-sim) home123: ~/Documents/astra-sim/extern/graph_frontend/chakra$ python3 -m chakra.et_converter.et_converter --input_type PyTorch --input_filename et_plus/dlrm_eg_0_plus.json --output_filename et_plus/dlrm_chakra.0.et --

Traceback (most recent call last):
  File "/miniconda3/envs/astra-sim/lib/python3.9/runpy.py", line 197, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/miniconda3/envs/astra-sim/lib/python3.9/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/miniconda3/envs/astra-sim/lib/python3.9/site-packages/chakra/et_converter/et_converter.py", line 9, in <module>
    from .text2chakra_converter import Text2ChakraConverter
  File "/miniconda3/envs/astra-sim/lib/python3.9/site-packages/chakra/et_converter/text2chakra_converter.py", line 8, in <module>
    from chakra.et_def.et_def_pb2 import (
  File "/miniconda3/envs/astra-sim/lib/python3.9/site-packages/chakra/et_def/et_def_pb2.py", line 9, in <module>
    from google.protobuf import runtime_version as _runtime_version
ImportError: cannot import name 'runtime_version' from 'google.protobuf' (/miniconda3/envs/astra-sim/lib/python3.9/site-packages/google/protobuf/__init__.py)

May I ask why this issue occurs? Could it be related to my version of protobuf not being compatible?

I noticed that the tutorial mentions the need for protobuf 4.x. I tried different versions like protobuf 4.21.6 and 4.24.4, with Python 3.9, and PyTorch 2.1.2 and 2.4.0, and the latest version of chakra, but I keep encountering the same error.

Start with the wiki's “Using Chakra Execution Trace and Kineto Traces Generated By PyTorch” section and the failing chakra.et_converter.et_converter entry point. Read text2chakra_converter.py and et_def_pb2.py around the google.protobuf import, then reproduce the command with the reported Python, PyTorch, and protobuf versions. Done means the converter command runs without the reported ImportError, with the compatible dependency requirement documented.