-
Notifications
You must be signed in to change notification settings - Fork 883
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Panic while fine-tuning with LORA #517
Comments
A few comments / questions:
There are some other things we/you can do to reduce memory:
I will plan to update the mlx lm lora to do bucketing + compilation with an option to checkpoint. But in the meantime you can experiment with those if you are comfortable digging in to the Python. |
Hay! thanks for the quick reply
|
@awni thanks a lot, just-re ran and it went smoooth.. great success! |
Well that's a surprise and a delight, great to hear!
So the term "checkpointing" is overloaded. In What I was referring to re checkpointing is gradient checkpointing which is a way to reduce memory use at the cost of computation. That's a totally different thing and is not currently used in any of our Lora examples. |
Hi,
Thanks everyone working on this awesome lib!
I've been trying to fine tune a 7b model on my M2 ULTRA 64GB machine.
After some time the machine panics.. I think it has something to do with memory..
RAM Usage: 55.8/64.0GB - swap:75.9/77.0GB
I don't want to compromise on the quality of the fine tune, can I do something to allow this to complete or pass some flag to have it run a bit slower but be able to finish?
The SSH session log:
The panic log:
The text was updated successfully, but these errors were encountered: