-
Notifications
You must be signed in to change notification settings - Fork 355
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
CI/CD setup #1137
CI/CD setup #1137
Conversation
…github.com/pytorch/TensorRT into circleci-editor/245/circleci-project-setup
@narendasan For documentation changes may add |
Hmm, ok, we still need some jobs to run to build the docs but we dont need everything to run i guess. |
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.
Code conforms to C++ style guidelines
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.
Code conforms to C++ style guidelines
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.
Code conforms to C++ style guidelines
…github.com/pytorch/TensorRT into circleci-editor/245/circleci-project-setup
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.
Code conforms to C++ style guidelines
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.
Code conforms to C++ style guidelines
@narendasan I addressed the comments before. If you are Ok with it, could we land it first and could you help to add the torchscript tests later? We can iterate based on this version as a beginning for our future test needs. |
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.
Code conforms to C++ style guidelines
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.
Just a minor thing about the name of the command other than that I am good to merge this and improve incrementally
py/setup.py
Outdated
|
||
def get_git_revision_short_hash() -> str: | ||
return subprocess.check_output(['git', 'rev-parse', '--short', 'HEAD']).decode('ascii').strip() | ||
|
||
if "--fx2trt-only" in sys.argv: |
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.
IMO we should call this --fx-only
or --fx-backend-only
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.
I think we should drop references to fx2trt for the most part
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.
Yeah exactly. fx-only is a much better name here.
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.
Code conforms to C++ style guidelines
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.
Code conforms to C++ style guidelines
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.
Code conforms to C++ style guidelines
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.
Code conforms to C++ style guidelines
Description
This is PR is not be ready for merge yet. There are some points we need to improve together with NV.
This PR is for setting up the CI/CD. Pytorch collaborates with circle CI to provide CI/CD. I am new to circle CI so I spent some efforts to read their doc and come up with our 1st version. Since we need to use GPU machine, we actually can only choose their machine executor with choice of 3 types of card P4,T4,V100
Their available GPU images come with cuda driver and their supporting libs installed. We can not use our own docker image as it is not supported (error msg below).
So based on the vanilla cuda environment, here are the steps that I set up in config.yaml (recipe to set up the CI/CD)
Overall, circle CI is good to use and can satisfy our needs for general CI/CD.
Pros:
Cons:
I want to discuss with NV folks in general, if we could deploy circle CI for our CI/CD. What else you think we can possibly improve?