[GPU] Fix FC shape agnostic kernel build failure issue. #22868
Merged
Conversation
This file contains 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
isanghao
reviewed
Feb 19, 2024
| if (node.is_type<fully_connected>()) { | ||
| if (allow_new_shape_infer) { | ||
| expected = format::get_default_format(node.get_input_layout(0).get_rank()); | ||
| node.set_preferred_input_fmt(0, expected); |
There was a problem hiding this comment.
Is it complicated to make a functional test case for such pattern? If not, I think it would be better to implement one..
There was a problem hiding this comment.
Added functional test.
There was a problem hiding this comment.
Could you please add a comment noting that plain input format is enforced because there is no available shape agnostic kernel supporting blocked format for now? Once we enable shape agnostic kernel for fsv->bfyx fc optimized kernel (i8), then we could relax the condition..
There was a problem hiding this comment.
Added the comment.
Signed-off-by: hyunback <[email protected]>
Signed-off-by: hyunback <[email protected]>
Signed-off-by: hyunback <[email protected]>
Signed-off-by: hyunback <[email protected]>
Signed-off-by: hyunback <[email protected]>
Signed-off-by: hyunback <[email protected]>
hyunback
force-pushed
the
fix_notebook_issue0216
branch
from
29ad8ef to
233c2b2
Compare
Signed-off-by: hyunback <[email protected]>
yeonbok
reviewed
Feb 21, 2024
| if (node.is_type<fully_connected>()) { | ||
| if (allow_new_shape_infer) { | ||
| // Plain input format is enforced because no available shape agnostic kernel supporting blocked format. | ||
| // TODO: will remove once enable shape agnostic kernel for fsv->bfyx FC optimized kernel(i8) |
There was a problem hiding this comment.
Suggested change
| // TODO: will remove once enable shape agnostic kernel for fsv->bfyx FC optimized kernel(i8) | |
| // TODO: The condition will be relaxed once more shape agnostic kernels for other formats are enabled (e.g., fsv->bfyx FC optimized kernel(i8))) |
Signed-off-by: hyunback <[email protected]>
yeonbok
approved these changes
Feb 26, 2024
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.
239-image-bind-quantize notebook fails in GPU.
Target model has
Conv - Fakequantize - FCpattern and FC was set fsv16 format. Currently FC shape agnostic kernel only support planar format, so enforce planar format in FC before supporting block format.Tickets: