how to create backends in different conda env ?
cc @architkulkarni Can you address the qusetion?
Thanks for the ping Sang! @Ti_n_Hoang For instructions, see Advanced Topics and Configurations — Ray v1.2.0 (or Core APIs — Ray v2.0.0.dev0 if you’re using the Ray nightly build).
Thanks for the info @architkulkarni
I’ve made some tests and I appreciate that ray serve must be installed in provided conda environment. I think it would be nice if models backends can be deployed with ray agnostic conda environments.