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.

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 dev packages for windows pytorch-cuda dependencies #1295

Merged
merged 5 commits into from
Feb 8, 2023
Merged
Show file tree
Hide file tree
Changes from 4 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions conda/pytorch-cuda/conda_build_config.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -2,3 +2,6 @@ version:
- 11.6
- 11.7
- 11.8
target_platform:
- win-64
- linux-64
23 changes: 17 additions & 6 deletions conda/pytorch-cuda/meta.yaml
Original file line number Diff line number Diff line change
@@ -1,11 +1,17 @@
# Package to manage cuda version in PyTorch
# Package to manage cuda version in PyTorch.
#
# Windows anaconda packages are packaged differently,
# All dlls are kept within *-dev packages hence we need
# include the dev packages for Windows see:
# https://github.com/pytorch/vision/issues/7185#issuecomment-1420002413
#
# Please note: Build number should be advanced with
# every deployment. After the deployment to production
# use following links to validate the correctness of
# deployment:
# https://conda.anaconda.org/pytorch/noarch/
# https://conda.anaconda.org/pytorch/noarch/repodata.json
{% set build = 3 %}
{% set build = 2 %}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it should be incremented upwards, so 4.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks like we are just moving faster then light, so time is spinning backward :D

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for review, we probably want to keep v2. Since in prod current latest one is v1:
https://anaconda.org/pytorch/pytorch-cuda/files

For nightlies and test I think we can keep v2.

{% set cuda_constraints=">=11.6,<11.7" %}
{% set libcufft_constraints=">=10.7.0.55,<10.7.2.50" %}
{% set libcublas_constraints=">=11.8.1.74,<11.10.1.25" %}
Expand Down Expand Up @@ -36,15 +42,11 @@ package:
version: {{ version }}
build:
number: {{ build }}
noarch: generic
requirements:
run:
- cuda-libraries {{ cuda_constraints }}
- cuda-nvtx {{ cuda_constraints }}
- libnvjpeg {{ libnvjpeg_constraints }}
# libnvjpeg-dev required see: https://github.com/pytorch/vision/issues/7185#issuecomment-1420002413
- libnvjpeg-dev {{ libnvjpeg_constraints }}
run_constrained:
- cuda-cudart {{ cuda_constraints }}
- cuda-cupti {{ cuda_constraints }}
- cuda-nvrtc {{ cuda_constraints }}
Expand All @@ -54,6 +56,15 @@ requirements:
- libcusolver {{ libcusolver_constraints }}
- libcusparse {{ libcusparse_constraints }}
- libnpp {{ libnpp_constraints }}
- cuda-libraries-dev {{ cuda_constraints }} # [win64]
- libnvjpeg-dev {{ libnvjpeg_constraints }} # [win64]
- cuda-cudart-dev {{ cuda_constraints }} # [win64]
- cuda-nvrtc-dev {{ cuda_constraints }} # [win64]
- libcufft-dev {{ libcufft_constraints }} # [win64]
- libcublas-dev {{ libcublas_constraints }} # [win64]
- libcusolver-dev {{ libcusolver_constraints }} # [win64]
- libcusparse-dev {{ libcusparse_constraints }} # [win64]
- libnpp-dev {{ libnpp_constraints }} # [win64]
test:
commands:
- echo "pytorch-cuda metapackage is created."
Expand Down