No Module named Deepspeed #370
Labels
No Label
bug
duplicate
enhancement
help wanted
insufficient info
invalid
news
not a bug
question
wontfix
No Milestone
No project
No Assignees
4 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: mrq/ai-voice-cloning#370
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Im trying to use the latest improvement "deepspeed". But so far its not working. When i turn this on in the settings,then i get this error
I think its not properly implemented..and also im so eager to try this out.
Getting error starting ai voice cloningto No Module named Deepspeedu can install it with
pip install deepspeed==0.10.2
Tested on WSL Ubuntu 20.04
Hi mate, Tried to install but got this error instead!
DeepSpeed isn't available on Windows natively without going through a nightmare of hoops to get it compiled and installed. The best bet is to have an environment under WSL2 but you're still under the overhead tax from operating under WSL2.
It's why I haven't included it as a dependency in requirements.txt, and why I marked it as disabled by default in the web UI.
As stated above, it can be a nightmare to get going. I run Windows and only need it for inference, so if that's the case for you, you can go to this issue here and use marcos pre-compiled wheels file: https://github.com/microsoft/DeepSpeed/issues/2588
Once you have it, it's a matter of pip installing it and it should work fine for inference. Though you need to be on Cu117 for this to work else you need to reinstall pytorch pacakges under 117.
Building the wheels yourself is a hell of a time sync and I don't recommend it. The author's instructions mostly work but I had to modify some of the steps to actually get it built.