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.

Sign up for GitHub

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

Jump to bottom

Fix checkfail in QuantizedConv2D with invalid input shape dimensions #63288

Merged
merged 2 commits into from
Mar 13, 2024
Merged

Fix checkfail in QuantizedConv2D with invalid input shape dimensions #63288

merged 2 commits into from
Mar 13, 2024

Conversation

SuryanarayanaY
Copy link
Collaborator

A checkfail can happen in raw_ops.QuantizedConv2D with invalid input shape i.e presence of 0 in input shape dimensions. Repro gist.

Fixes #59927 .

@SuryanarayanaY
Fix checkfail in QuantizedConv2D with invalid input shape dimensions
07e29eb 
A checkfail can happen in QuantizedConv2D with invalid input shape i.e presence of 0 in input shape dimensions. Repro gist.

Fixes #59927 .
@google-ml-butler google-ml-butler bot added the size:XS CL Change Size: Extra Small label Mar 8, 2024
@google-ml-butler google-ml-butler bot added kokoro:force-run Tests on submitted change ready to pull PR ready for merge process labels Mar 8, 2024
@kokoro-team kokoro-team removed the kokoro:force-run Tests on submitted change label Mar 8, 2024
@gbaned gbaned added this to Assigned Reviewer in PR Queue via automation Mar 8, 2024
@gbaned gbaned added the comp:core issues related to core part of tensorflow label Mar 8, 2024
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@golechwierowicz @tensorflower-gardener
[XLA:GPU] Strengthen abs/rel error in softmax triton emitter.
e3f6a80 
FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 615401428
@copybara-service copybara-service bot mentioned this pull request Mar 13, 2024
Closed
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@yangustc07 @tensorflower-gardener
#tf-data Support global shuffle for the Tensor slices dataset.
052f5ef 
FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 615218481
@copybara-service copybara-service bot mentioned this pull request Mar 13, 2024
Merged
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@yangustc07 @tensorflower-gardener
#tf-data-service Merge duplicate method `GetIteratorContextWithIndexM…
c87c69c 
…apper`.

Also add a simple class to select the right `IteratorContext` when index
shuffling is needed.

FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 615107953
@copybara-service copybara-service bot mentioned this pull request Mar 12, 2024
Merged
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@tensorflower-gardener
Add the options to require complete AOT autotuning results
6834c90 
This cl adds an XLA flag `--require_complete_aot_autotune_results` that requires complete AOT autotuning results inside XLA, so that a `NotFound` will be raised when there is cache miss and JIT autotuning is needed. This can help users detect JIT autotuning happening inside XLA.

FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 613961149
@copybara-service copybara-service bot mentioned this pull request Mar 13, 2024
Merged
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@klucke @tensorflower-gardener
Remove DeviceOptions from StreamExecutorConfig.
ce97679 
It's only used by CudaPlatform, so do the equivalent work there.

FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 614837803
@copybara-service copybara-service bot mentioned this pull request Mar 13, 2024
Closed
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@yangustc07 @tensorflower-gardener
#tf-data Support global shuffle for the Tensor slices dataset.
605a80e 
FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 615218481
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@yangustc07 @tensorflower-gardener
#tf-data-service Merge duplicate method `GetIteratorContextWithIndexM…
9f06e5f 
…apper`.

Also add a simple class to select the right `IteratorContext` when index
shuffling is needed.

FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 615107953
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@yashk2810 @tensorflower-gardener
Lower memory_kind on inputs/outputs to mhlo.memory_kind and then tr…
22e9602 
…anslate that to entry_computation_layout attached to HLO.

This is very similar to how layouts are plumbed from stablehlo to entry_computation_layout.

FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 606414264
@copybara-service copybara-service bot mentioned this pull request Mar 13, 2024
Closed
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@tensorflower-gardener
Add the options to require complete AOT autotuning results
f50c219 
This cl adds an XLA flag `--require_complete_aot_autotune_results` that requires complete AOT autotuning results inside XLA, so that a `NotFound` will be raised when there is cache miss and JIT autotuning is needed. This can help users detect JIT autotuning happening inside XLA.

FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 613961149
copybara-service bot pushed a commit that referenced this pull request Mar 13, 2024
@jcai19 @tensorflower-gardener
Rename more functions
695fbdf 
Rename TPU to Replicated in more functions in TF2XLA MLIR phase 1 bridge.

FUTURE_COPYBARA_INTEGRATE_REVIEW=#63288 from tensorflow:Surya_QuantizedConv2D_checkfail_fix bba7c37
PiperOrigin-RevId: 611193698
@copybara-service copybara-service bot mentioned this pull request Mar 13, 2024
Merged
@copybara-service copybara-service bot merged commit 0b6c03c into master Mar 13, 2024
13 of 15 checks passed
PR Queue automation moved this from Assigned Reviewer to Merged Mar 13, 2024
@copybara-service copybara-service bot deleted the Surya_QuantizedConv2D_checkfail_fix branch March 13, 2024 17:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@mihaimaruseac mihaimaruseac mihaimaruseac approved these changes

Assignees

@gbaned gbaned

Labels
comp:core issues related to core part of tensorflow ready to pull PR ready for merge process size:XS CL Change Size: Extra Small
Projects
PR Queue
  
Merged
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

A crash due to check-fail can be triggered in QuantizedConv2D
4 participants
@SuryanarayanaY @mihaimaruseac @kokoro-team @gbaned