Fleet can't find Python interpreter


Upon trying to select a Python interpreter, fleet states that there are none.  I won't detect any versions of Python currently installed, or any conda or venv environments on macOS.  I've tried manually selecting the interpreter file but upon opening it, Fleet does nothing and continues to say there are no interpreters.  Any idea on what might be causing this?

Official comment

Please make sure you have the Smart mode enabled. Then configure the interpreter for the project, please see Configure a Python interpreter.

I ran into this situation too, click "add python interpreter" and nothing happens, in settings there is a configuration about python interpreter but it only navigates me to "settings.json" while there is no docs for toolchains configuration


Did you try creating a venv or .venv directory within your project directory?  When I did that on Linux, and turned on Smart Mode, Fleet 1.9.237 noticed that venv and started using it, even though it didn't notice other pythons on my machine.


Andrey Dernov I've already followed those instructions and it is not working.  Gartic I just tried using venv and didn't have any luck. 


I had success when I pointed the python interpreter search to my conda environment (/home/user_name/anaconda3/envs/GNN/bin) and then selected python3. Replace user_name and GNN.


maybe you should support mutiple kinds of interpreter like pycharm does. conda, venv, ssh, wsl, docker, etc.


ZhangHC yes, we plan to improve conda and venv Python interpreter detection. Please vote for and follow these corresponding issues:

FL-14903 Unable to add Python Interpreter from .venv

FL-14881 Support for conda environments (and venv)

venv is supported for now only with standard names.


As for ssh and Docker remote interpreters - it is supported to work with remote interpreters but Fleet defines another approach: instead of ssh/wsl/Docker remote interpreters, Fleet Server should be spawned inside the container (Docker, WSL, any other remote server) within the project code inside that container. Then you can connect to it and develop and run remote code.

For Docker, you can also use Inject into container action tat will automatically install Fleet there:

Then, Connect to Docker... action might be used.


my project involves a gan module and it doesn't detect torch interpreter


Please sign in to leave a comment.