Repository navigation
Conversation
|
This PR changes source code for the following kernel(s): If you have the kernel-bot build permissions, then trigger |
drbh
marked this pull request as draft
November 10, 2026 19:08
Collaborator
Author
|
waiting on #1328 to run build |
Signed-off-by: David Holtz <david.richard.holtz@gmail.com>
Collaborator
Author
|
/kernel-bot build cv-utils |
|
Build request processed. Command: Dispatched (1):
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
this pr adds a rocm build for
cv-utilssame change as causal-conv1d in #1319: the
rocmbackend,rocm = "2.10"stable abi and acpp-rocmkernel with the same sources andactivationarchs, andstable_utils.hpicks the hip runtime underUSE_ROCM. the kernels have no warp intrinsics and nms already uses 64 bit masks, so no kernel changes. the triton functions from #1182 run on rocm as isbuilt all rocm variants (
rocm70,rocm71,rocm714,rocm72,rocm100) and ran the tests on mi300x (gfx942) with torch2.10.0+rocm7.0,2.12.1+rocm7.1and2.14.1+rocm7.2, 136 pass on each (4 skipped, they need two gpus).generic_nmskeeps the same boxes astorchvision.ops.nmson rocm at 1k, 5k and 20k boxes and is within 4% of its speednote: this lands in v2, sam3 video still loads
version=1so transformers needs to bump that before it picks up the rocm build. no version bump here since this only adds a backend*keeping as draft until ci support rocm