Protobuf version issue

[Usmaniatech]

How do you solve this issue @kotogasy

[handria-ntoanina]

@Usmaniatech , try to create a new environment using anaconda and following the instruction from readme.me under requirement section

[Usmaniatech]

I have followed the instructions and I do not find any python folder after installing tensorflow gpu

Install the dependencies under the folder python/ cd python pip install .

What dependencies does this folder installs.

[handria-ntoanina]

I have just added the python folder. You can pull the repository and follow the instructions

[handria-ntoanina]

@Usmaniatech it is better to create a fresh new environment. Then, did you follow these: Perform a minimal install of OpenAI gym

• if using Windows, -- download swig for windows and add it the PATH of windows -- install Microsoft Visual C++ Build Tools
• then run these commands -- pip install gym -- pip install gym[classic_control] -- pip install gym[box2d]

If you are using windows and did not install swig and build tools C++, you won't be able to install box2d

[Usmaniatech]

When we create build of the environment we have to check the controlbox for the training in unity. If I check it, that needs mlagents. If I dont check it, then it dont load the environment.exe file. Do we have to install dependencies in mlagents and mlagents-env folder. Because it creates conflicts with unityagents. So my question is, do we need mlagents for this example to work or it works without that.

[handria-ntoanina]

@Usmaniatech , I am just from following the entire requirements steps from scratch. I have added another step to install pytorch. Afterwards, I run the notebook and it ran very well.

[Usmaniatech]

Which version of ml-agents you are using ? @kotogasy

[handria-ntoanina]

@Usmaniatech just use the one prebuilt environment