Skip to content
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.

Sign up for GitHub

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

Jump to bottom

✨[Feature] PyPi binaries #1402

Closed
msaroufim opened this issue Oct 13, 2022 · 11 comments · Fixed by #1677
Closed

✨[Feature] PyPi binaries #1402

msaroufim opened this issue Oct 13, 2022 · 11 comments · Fixed by #1677
Assignees
@narendasan
Labels
channel: pip issues related to pip install feature request New feature or request release: v1.3 Tagged to be included in v1.3

Comments

@msaroufim
Copy link
Member

Is your feature request related to a problem? Please describe.
Having to use a custom index is less than ideal if I'd like to for example have this library be part of extra dependencies in my setup.py

Describe the solution you'd like
pip install torch-tensorrt
@msaroufim msaroufim added the feature request New feature or request label Oct 13, 2022
@narendasan
Copy link
Collaborator

We already have the name reserved on pypi but we cannot currently upload builds since some of our dependencies are not compliant with the manylinux spec. This may change in the future and when it does we will push packages there.

@msaroufim
Copy link
Member Author

@ncomly-nvidia do you think this is happening this year? I was hoping to create a seamless installation experience for my users in torchserve pytorch/serve#1857

@ncomly-nvidia ncomly-nvidia added the release: v1.3 Tagged to be included in v1.3 label Oct 18, 2022
@msaroufim
Copy link
Member Author

Thank you @ncomly-nvidia !

@ncomly-nvidia
Copy link
Contributor

@msaroufim sure thing. We have been gated by TensorRT's availability on PyPI. It will be available early Nov. so torch-trt should be able to get this up shortly there after. By the end of the year seems reasonable.

Looking forward to your working in torch serve!

@ncomly-nvidia ncomly-nvidia added the channel: pip issues related to pip install label Oct 19, 2022
@ychfan
Copy link

ychfan commented Oct 20, 2022

Is TensorRT available here?

@ncomly-nvidia
Copy link
Contributor

Hi @ychfan yes it is, however, the package is undergoing changes internally which will be finalized in the next TRT release in Nov. Torch-TRT will pickup this package after the changes.

@narendasan narendasan mentioned this issue Dec 5, 2022
Merged
@Christina-Young-NVIDIA
Copy link
Collaborator

Expected completion by EOW.

@Christina-Young-NVIDIA
Copy link
Collaborator

Waiting on CI fixes. Naren following up.

@Christina-Young-NVIDIA
Copy link
Collaborator

Now unblocked.

@narendasan narendasan mentioned this issue Feb 15, 2023
Merged
7 tasks
@ncomly-nvidia
Copy link
Contributor

@narendasan can we please post a package to PyPI. I know there are still outstanding improvements we are considering getting in - can we post v1.3.0 & update to v1.3.1 if needed?

@ncomly-nvidia ncomly-nvidia reopened this Mar 16, 2023
@narendasan narendasan assigned narendasan and unassigned peri044 Mar 21, 2023
@narendasan
Copy link
Collaborator

Binaries are uploaded

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@narendasan narendasan

Labels
channel: pip issues related to pip install feature request New feature or request release: v1.3 Tagged to be included in v1.3
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Cherry-pick manylinux compatible builds into main pytorch/TensorRT
6 participants
@narendasan @ychfan @msaroufim @peri044 @ncomly-nvidia @Christina-Young-NVIDIA