Skip to content

Conversation

@mgehre-amd
Copy link
Collaborator

No description provided.

zjgarvey and others added 6 commits September 17, 2024 16:01
Previously, if the value was absent, this conversion was creating a
dense resource of value 0 with shape equal to the result shape, then
later re-extracting a splat value. This only works if the shape is
statically known, and even when the shape is known, this is completely
unnecessary since the value's shape should be `[1]` and not the result
shape.

This patch simply sets the `splatvalue` to a `torch.constant.float 0.0`
when the onnx op's `value` attr is absent, and adds `nullptr` checks to
the subsequent conditionals to avoid them in the case where an `attr` is
not given.

Addresses <nod-ai/AMD-SHARK-ModelDev#831>.
Set PyTorch and TorchVision version to nightly release 2024-09-16.

Signed-Off By: Vivek Khandelwal <[email protected]>
@mgehre-amd mgehre-amd requested a review from jorickert December 12, 2024 07:51
Base automatically changed from bump_to_14ef05a2 to feature/backport_ea1_ops December 13, 2024 14:59
@mgehre-amd mgehre-amd merged commit d03e3fa into feature/backport_ea1_ops Dec 13, 2024
4 checks passed
@mgehre-amd mgehre-amd deleted the bump_to_3f46348e branch December 13, 2024 14:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants