release-21-branch numpy can't install successful.

[Feipian]

When you install mlagent-env using python virtual environment according to the install instruction document, you will encounter numpy can't install, This is because python 3.10.12 does not support the numpy version of setup.py in the ml-agents-envs folder, so setup.py needs to be changed

import os
import sys
from setuptools import setup, find_packages
from setuptools.command.install import install
import mlagents_envs

VERSION = mlagents_envs.__version__
EXPECTED_TAG = mlagents_envs.__release_tag__

here = os.path.abspath(os.path.dirname(__file__))


class VerifyVersionCommand(install):
"" "
Custom command to verify that the git tag is the expected one for the release.
Originally based on https://circleci.com/blog/continuously-deploying-python-packages-to-pypi-with-circleci/
This differs slightly because our tags and versions are different.
"" "

description = "verify that the git tag matches our version"

def run(self):
tag = os.getenv("GITHUB_REF", "NO GITHUB TAG!" ).replace("refs/tags/", "")

if tag ! = EXPECTED_TAG:
info = "Git tag: {} does not match the expected tag of this app: {}".format(
tag, EXPECTED_TAG
)
sys.exit(info)


# Get the long description from the README file
with open(os.path.join(here, "README.md"), encoding="utf-8") as f:
long_description = f.read()

setup(
name="mlagents_envs",
version=VERSION,
description="Unity Machine Learning Agents Interface",
long_description=long_description,
long_description_content_type="text/markdown",
url="https://github.com/Unity-Technologies/ml-agents",
author="Unity Technologies",
author_email="ML-Agents@unity3d.com",
classifiers=[
"Intended Audience :: Developers",
"Topic :: Scientific/Engineering :: Artificial Intelligence",
"License :: OSI Approved :: Apache Software License",
Programming Language :: Python :: 3.10",
].
packages=find_packages(
exclude=["*.tests", "*.tests.*", "tests.*", "tests", "colabs", "*.ipynb"]
),
zip_safe=False,
install_requires=[
"cloudpickle",
"grpcio> = 1.11.0, & lt; = 1.48.2 ",
"Pillow> = 2 ",
"protobuf> = 3.6, & lt; 3.20 ",
"pyyaml> = 3.1.0 ",
"gym> = 0.21.0 ",
"Pettingzoo = = 1.15.0",
"numpy==1.22.4", # change the numpy version, because default version will display 'can't successful install numpy'
"filelock> = 3.4.0 ",
].
python_requires="> = 3.10.1, & lt; = 3.10.12 ",
# TODO: Remove this once mypy stops having spurious setuptools issues.
cmdclass={"verify": VerifyVersionCommand},  # type: ignore
)

Can run

[AbhiG7]

Hey, I've been having the same issue as you so I tried changing the numpy version in the file like you showed and it allowed me to install the packages but now I get another error and I was wondering if you were getting it too. I get an error that says:
Warning: Failed to initialize NumPy: module compiled against API version 0x10 but this version of numpy is 0xf (function operator ()).
Have you gotten this before and do you know how to fix this?

[Feipian]

Maybe you come cross this problem and try:https://stackoverflow.com/questions/71986643/userwarning-failed-to-initialize-numpy-module-compiled-against-api-version-0xf

[Feipian]

Hey, I've been having the same issue as you so I tried changing the numpy version in the file like you showed and it allowed me to install the packages but now I get another error and I was wondering if you were getting it too. I get an error that says: Warning: Failed to initialize NumPy: module compiled against API version 0x10 but this version of numpy is 0xf (function operator ()). Have you gotten this before and do you know how to fix this?

General flow:

• Install Microsoft visual C++
• Install Visual Studio with Unity and C++
• clone mlagents
• Reboot
• Creating a virtual environment python==3.10.12 (anaconda)
• installation pytorch (according to mlagents document version) : pip3 install torch ~ = 1.13.1 - f https://download.pytorch.org/whl/torch_stable.html (follow mlagent document)
• Change setup.py in /ml-agents-envs (it may not support numpy)
• I changed it to "numpy==1.22.4", (setup.py content) can't use numpy Most likely mlagents default numpy version is no longer supported by python3.10.12
• After that it's probably unity to install the mlagent plugin
• You can type 'mlagens-learn' in python with mlagent enabled in CMD and if it pops up, it's successful

[Feipian]

Hey, I've been having the same issue as you so I tried changing the numpy version in the file like you showed and it allowed me to install the packages but now I get another error and I was wondering if you were getting it too. I get an error that says: Warning: Failed to initialize NumPy: module compiled against API version 0x10 but this version of numpy is 0xf (function operator ()). Have you gotten this before and do you know how to fix this?

I change numpy version to 1.22.4 because it support python3.10.12

[AbhiG7]

I followed those steps but met with many issues with the numpy version. I ended up finding the right solution for me, and it was to change the numpy verison in the setup.py file in the ml-agents-envs folder to "numpy>=1.21.2,<1.24.0" rather than the other suggested ways. It's shown in this screenshot:

which is taken from this branch of the ml-agents repo: https://github.com/Unity-Technologies/ml-agents/blob/develop/ml-agents-envs/setup.py.

Thanks for the help, though!

[ICOnce]

I am having a problem that causes this error when running mlagents-learn --help after having changed the numpy version and successfully gotten it to install

[ICOnce]

I am having a problem that causes this error when running mlagents-learn --help after having changed the numpy version and successfully gotten it to install

I seem to have solved, that issue, but now it gives a numpy error upon typing mlagents-learn

[ashwinsnambiar]

I had faced a similar issue. It seemed to be caused by an incorrect numpy version. I removed the existing numpy and mlagents packages from the virtual env and reinstalled everything by using the setup.py from a branch release.
Please check the following link for the same.
#6008 (comment)

[IvanVeridian]

I had this issue as well and fixed it by following @ashwinsnambiar recommendation of following @onurkurum #6008 (comment)

"Workaround:
Changed
"numpy==1.21.2",
to
"numpy==1.23.3",

in the file path-to-ml-agents\ml-agents-envs\setup.py"

[ICOnce]

I ended up fixing all the issues, here is a document about how I fixed everything
MLAgents setup guide.pdf

[Feipian]

I ended up fixing all the issues, here is a document about how I fixed everything MLAgents setup guide.pdf

Very detailed teaching, thank you!!

[ICOnce]

I ended up fixing all the issues, here is a document about how I fixed everything MLAgents setup guide.pdf

Very detailed teaching, thank you!!

NP!