-
Notifications
You must be signed in to change notification settings - Fork 365
enable back jetpack build #3720
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
base: release/2.8
Are you sure you want to change the base?
Conversation
1) no longer publish jetpack build to pytorch nightly 2) publish jetpack build to pypi.jetson-ai-lab only for each release 3) no longer use --plat-name=linux_tegra_aarch64 for jetson wheel user need to download the wheel from pypi/pytorch index for Thor and download the wheel from pypi.jetson-ai-lab for Orin
@@ -61,6 +61,8 @@ System Preparation | |||
sudo cp -a libcusparse_lt-linux-sbsa-0.5.2.1-archive/lib/* /usr/local/cuda/lib64/ | |||
|
|||
Building Torch-TensorRT | |||
|
|||
Since now Thor is using the SBSA build, we only build the jetpack wheel for Orin. |
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.
Remove this until Thor is public
axis, | ||
) | ||
else: | ||
if Version(metadata.version("tensorrt")) >= Version("10.8.0"): |
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.
Is this how we do version checks elsewhere in the library?
|
||
from packaging.version import Version | ||
|
||
if Version(metadata.version("tensorrt")) >= Version("10.8.0"): |
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.
Should we create a macro or utility function for this?
@@ -79,7 +79,9 @@ def load_dep_info(): | |||
dir_path = os.path.join(str(get_root_dir()), "py") | |||
|
|||
IS_AARCH64 = platform.uname().processor == "aarch64" | |||
IS_JETPACK = True if "tegra" in platform.uname().release else False | |||
# we do not auto detect if it is a tegra platform to enable JETPACK build | |||
# since we have thor using sbsa build, we need to check if the build has --jetpack flag |
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.
Remove mentions of thor until jp7 release
|
||
"tensorrt>=10.12.0,<10.13.0; platform_machine != 'aarch64' or (platform_machine == 'aarch64' and 'tegra' not in platform_release)", | ||
"tensorrt-cu12>=10.12.0,<10.13.0; platform_machine != 'aarch64' or (platform_machine == 'aarch64' and 'tegra' not in platform_release)", | ||
"tensorrt-cu12-bindings>=10.12.0,<10.13.0; platform_machine != 'aarch64' or (platform_machine == 'aarch64' and 'tegra' not in platform_release)", | ||
"tensorrt-cu12-libs>=10.12.0,<10.13.0; platform_machine != 'aarch64' or (platform_machine == 'aarch64' and 'tegra' not in platform_release)", | ||
|
||
"tensorrt>=10.3.0,<10.4.0;platform_machine == 'aarch64' and 'tegra' in platform_release", | ||
"tensorrt-cu12>=10.3.0,<10.4.0; platform_machine == 'aarch64' and 'tegra' in platform_release", |
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.
Why dont we need these?
only publish jetpack build to pypi.jetson-ai-lab for each release manually
user need to download the wheel from pypi/pytorch index for Thor
and download the wheel from pypi.jetson-ai-lab for Orin
Description
Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context. List any dependencies that are required for this change.
Fixes # (issue)
Type of change
Please delete options that are not relevant and/or add your own.
Checklist: