-
Notifications
You must be signed in to change notification settings - Fork 24
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
Fixes #261 partially #426
Fixes #261 partially #426
Conversation
MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅ |
Thank you @Oseltamivir for your changes. Can you please tell if you are from an MLCommons member organization? Can you please sign the CLA? |
@arjunsuresh Sorry for late reply, I just signed the CLA Not sure how to get bot to recheck |
Thank you @Oseltamivir The CLA test now passes. There is a problem for
|
Thanks @arjunsuresh Yea, I think your regex will be more suitable. Also, I had some other problems finding versions for: And some other packages have been tested on ubuntu only (Those with Also to add onto documentation if needed, I used https://regexr.com/ for regex testing and https://regex-generator.olafneumann.org for generating complex regexes |
@Oseltamivir yes, thats correct. Most times the version detection shouldn't change depending on the package manager but we still need to test. Would you like to join the weekly 30 minutes CM sync to discuss on this? It is at GMT 6:05pm Tuesdays. |
I apologise but I dont think I can make it. I'm in GMT+8 which is around 3am... maybe I'll try next week. |
No worries @Oseltamivir Let me know a good time and we can sync separately. Usually GMT 9-12 works best for me. |
I think next week's CM sync should be fine. I just need to plan ahead. Sleeping nowadays at 3am anyway ¯_(ツ)_/¯ |
No worries. Good night! :) |
@Oseltamivir Your PR was a big motivation for us to add the |
@arjunsuresh Awesome! Also, if you celebrate it, Happy Diwali! Let me know if there are any regexes/tasks you need me to do regarding this aspect. Meanwhile, I will try to resolve issues and see if i can contribute to this repo and mlcommons/inference - my current focus is attempting to make the default reference implementation use all GPUs on a single node. |
Thank you @Oseltamivir It'll be great if this PR is fixed and merged as it adds a very useful feature. "reference implementation use all GPUs on a single node" This is very useful as and some of the reference implementations like for llama2-70b already use all available GPUs on a node. But most of them don't. |
Sure. No prob. I'll try to slowly go thru the failing tests |
Got very frustrated over a specific library
Added a specific clause in the Got a feeling the errors for I will move onto check the other regexes soon. Currently wondering why |
Oh. That's an interesting find. Thank you for digging into it. I think it's better to set the below variable in
If |
Current issues: For macOS build:
For ubuntu: Not sure why this errors... on a numa supported device it shld work. Tested to be working on docker from runner and ubuntu bare metal. |
@Oseltamivir I guess I have fixed the errors on macos - the problem is |
Add Version RE for g++-11
Use berkeley link for imagenet-aux by default
Hi @Oseltamivir I removed the version detection ones which are failing and those which reply on |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you for the changes
The changes so far has been tested on a ubuntu 22.04 docker image.
Some utilities that are not yet tested:
Packages:
libmkl-dev
linux-tools