-
Notifications
You must be signed in to change notification settings - Fork 226
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
Add quant workflow to validate.sh #229
Conversation
e793a4a
to
7418f2a
Compare
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.
Please ensure at least as much green as before landing!
float16 w/ quant is generating garbage output on all platforms, for example, https://github.com/pytorch/torchchat/actions/runs/8716236562/job/23909449211?pr=229#step:7:2032 I will disable float16 for now. Wait, I don't see export.py and generate.py can take explicit |
7418f2a
to
16624eb
Compare
16624eb
to
e313e34
Compare
To ensure no new failures, trigger a periodic run on large models: https://github.com/pytorch/torchchat/actions/runs/8717420375 |
Segment fault with AOTI + quant int4 on linux x86_64 is an known issue iirc. |
The segfault with int4 groupwise quant in test-cpu-aoti x86_64 is an known issue, other than that no new failures. So I will ignore it and start merging. |
Merge quant flow (w/ default dtype) to pull.yml/periodic.yml, for both cpu and cuda
Remove dup yml files, compile.yml, and compile_t4.yml