Question Need help with MLAgents

Hi. I am trying to get the ML agents examples scene setup and ran into something - it doesn't work. I have zero python/ml history so I was just following a tutorial. Can someone kindly tell me what I am doing wrong?

So, here is what I did

• Downloaded and installed Anaconda, created env

conda create -n mlagents python=3.8
conda activate mlagents
pip3 install torch -f https://download.pytorch.org/whl/torch_stable.html
pip3 install -e ./ml-agents-envs
pip3 install -e ./ml-agents​

• Started a new project on 2021.3.18, added Newtonsoft Json package
• Added package in disk Release 19 ML package
• Copied the downloaded package (the ML-Agents folder)
• Then when I enter this in Anaconda (to begin training), mlagents-learn config/ppo/PushBlock.yaml --run-id=push_block_test_01

I get this, I've tried Release 20, different versions of Unity, python=3.9
Can someone kindly help?

(mlagents19) C:\Users\temp\Documents\mlAgents19\ml-agents-release_19>mlagents-learn config/ppo/PushBlock.yaml --run-id=push_block_test_01
Traceback (most recent call last):
File "C:\Users\temp\anaconda3\envs\mlagents19\Scripts\mlagents-learn-script.py", line 33, in <module>
sys.exit(load_entry_point('mlagents', 'console_scripts', 'mlagents-learn')())
File "C:\Users\temp\anaconda3\envs\mlagents19\Scripts\mlagents-learn-script.py", line 25, in importlib_load_entry_point
return next(matches).load()
File "C:\Users\temp\anaconda3\envs\mlagents19\lib\importlib\metadata.py", line 77, in load
module = import_module(match.group('module'))
File "C:\Users\temp\anaconda3\envs\mlagents19\lib\importlib\__init__.py", line 127, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "<frozen importlib._bootstrap>", line 1014, in _gcd_import
File "<frozen importlib._bootstrap>", line 991, in _find_and_load
File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
File "<frozen importlib._bootstrap_external>", line 843, in exec_module
File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents\mlagents\trainers\learn.py", line 2, in <module>
from mlagents import torch_utils
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents\mlagents\torch_utils\__init__.py", line 1, in <module>
from mlagents.torch_utils.torch import torch as torch # noqa
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents\mlagents\torch_utils\torch.py", line 6, in <module>
from mlagents.trainers.settings import TorchSettings
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents\mlagents\trainers\settings.py", line 25, in <module>
from mlagents.trainers.cli_utils import StoreConfigFile, DetectDefault, parser
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents\mlagents\trainers\cli_utils.py", line 5, in <module>
from mlagents_envs.environment import UnityEnvironment
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents-envs\mlagents_envs\environment.py", line 12, in <module>
from mlagents_envs.side_channel.side_channel import SideChannel
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents-envs\mlagents_envs\side_channel\__init__.py", line 5, in <module>
from mlagents_envs.side_channel.default_training_analytics_side_channel import ( # noqa
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents-envs\mlagents_envs\side_channel\default_training_analytics_side_channel.py", line 7, in <module>
from mlagents_envs.communicator_objects.training_analytics_pb2 import (
File "c:\users\temp\documents\mlagents19\ml-agents-release_19\ml-agents-envs\mlagents_envs\communicator_objects\training_analytics_pb2.py", line 35, in <module>
_descriptor.FieldDescriptor(
File "C:\Users\temp\anaconda3\envs\mlagents19\lib\site-packages\google\protobuf\descriptor.py", line 561, in __new__
_message.Message._CheckCalledFromGeneratedFile()
TypeError: Descriptors cannot not be created directly.
If this call came from a _pb2.py file, your generated code is out of date and must be regenerated with protoc >= 3.19.0.
If you cannot immediately regenerate your protos, some other possible workarounds are:
1. Downgrade the protobuf package to 3.20.x or lower.
2. Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python (but this will use pure-Python parsing and will be much slower).

More information: https://developers.google.com/protocol-buffers/docs/news/2022-05-06#python-updates

 

Sorry if this is noob question. I just don't know what to do
If anyone can help, it will be really appreciated. I have zero idea what is going on or what to do.

 

downgrade protobuf

Code (csharp):

1.  
2. pip install protobuf==3.20.3
3.  

(I suspect that you installed the packages in the reverse order, and creating a fresh env, and installgin simply

Code (csharp):

1.  
2. conda create -n mlagents python=3.8
3. conda activate mlagents
4. pip3 install -e ./ml-agents
5.  

.... will work better actually... )

 

Thank you! I will give it a try!

 

It worked!!!!!!!

just one last question. If I want to start again, let's say after a pc reboot. I am guessing
conda activate mlagents

and then

mlagents-learn config/ppo/PushBlock.yaml --run-id=push_block_test_01

to begin training again, correct? (no need to install things again)

Sorry if this is a noob question.

 

That sounds correct, yes

 

Also, if I am starting a fresh project, so 1 mlagents sample at the moment, and 2 or three new projects, do I need to create a separate environment and all files (so repeating the whole process for each project) or can I just leave the python files and create multiple projects? (provided they all use the same versions)

 

Reusing the python environment is fine.