Suggestion: update numpy version in ml-agents-envs/setup.py

[xyz2022]

Note: problem and solution contained within

context

Following this guide: https://github.com/Unity-Technologies/ml-agents/blob/develop/docs/Installation.md,
Which points to this repo: git clone --branch release_21 https://github.com/Unity-Technologies/ml-agents.git,
line 63 of file ml-agents/ml-agents-envs
/setup.py: "numpy==1.21.2"

problem

` Building wheel for numpy (pyproject.toml) ... error
error: subprocess-exited-with-error

× Building wheel for numpy (pyproject.toml) did not run successfully.
│ exit code: 1
╰─> [214 lines of output]
setup.py:63: RuntimeWarning: NumPy 1.21.2 may not yet support Python 3.10.`

So try numpy==1.22.1 and it installs fine but new error when its time to run mlagents-learn: Failed to initialize NumPy: module compiled against API version 0x10 but this version of numpy is 0xe

Solution

Finally, pip install numpy==1.23.1 and everything works fine. No errors on install or on mlagents-learn

Conclusion:

For users installing the following this guide verbatim: https://github.com/Unity-Technologies/ml-agents/blob/develop/docs/Installation.md

The version of numpy in setup.py (in ml-agents-envs/setup.py and ml-agents/setup.py) should be changed to:
"numpy==1.23.1"

[xyz2022]

Environment info:
(mlagents) PS C:\ML2\ml-agents> python --version
Python 3.10.13
(mlagents) PS C:\ML2\ml-agents> pip list
Package Version

---

absl-py 2.0.0
attrs 23.1.0
cachetools 5.3.2
cattrs 1.5.0
certifi 2023.7.22
charset-normalizer 3.3.1
cloudpickle 3.0.0
colorama 0.4.6
filelock 3.13.0
fsspec 2023.10.0
google-auth 2.23.3
google-auth-oauthlib 1.1.0
grpcio 1.48.2
gym 0.26.2
gym-notices 0.0.8
h5py 3.10.0
huggingface-hub 0.18.0
idna 3.4
Markdown 3.5
MarkupSafe 2.1.3
mlagents 1.0.0
mlagents-envs 1.0.0
numpy 1.23.1
oauthlib 3.2.2
onnx 1.12.0
packaging 23.2
PettingZoo 1.15.0
Pillow 10.1.0
pip 23.3
protobuf 3.19.6
pyasn1 0.5.0
pyasn1-modules 0.3.0
pypiwin32 223
pywin32 306
PyYAML 6.0.1
requests 2.31.0
requests-oauthlib 1.3.1
rsa 4.9
setuptools 68.0.0
six 1.16.0
tensorboard 2.15.0
tensorboard-data-server 0.7.2
torch 1.13.1+cu117
tqdm 4.66.1
typing_extensions 4.8.0
urllib3 2.0.7
Werkzeug 3.0.1
wheel 0.41.2 '

[xyz2022]

Additional note:
After posting this issue, I've seen in commit notes and PRs that python version must be 3.10.12
The doc's however, say "python 3.10.12 or higher".
Finally, the doc's say: run conda create -n mlagents python=3.10 && conda activate mlagents, this results in python 3.10.13 being installed. Which is higher than 3.10.12.
To be honest I think the maintainer should simply update the numpy version in setup.py to avoid all the headaches caused by the the doc's and numpy errors.
Can confirm: Windows + conda + python 3.10.13 + numpy 1.23.1 is a working combination.

[miguelalonsojr]

This has already been patched on develop and will be available next release. #5997