I do work with audio that requires librosa. While ChatGPT’s base environment has librosa installed, the system’s NumPy and Numba versions are not compatible. This makes it impossible to use librosa as-is.
However, I realize that, in some cases, I am able to install whl files. Sometimes the system fails the install, though other times it is successful. While this is annoying to repeat every new chat, are there any other suggestions on how to setup an environment within ChatGPT’s limitations?
Is it possible for enterprise users to specify their own environment or system?
Thank you for your help.