-
Notifications
You must be signed in to change notification settings - Fork 76.1k
Pull requests: tensorflow/tensorflow
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Update cudnn codegen tests to backend_config syntax from cudnn_frontend 1.27.0.
#126089
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
[XLA:GPU] Add an HLO pass to alias in-place outputs for Triton fusions and cuBLASLt matmuls.
#126087
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Register missing AdjustContrastv2 gradient
awaiting review
Pull request awaiting review
size:M
CL Change Size: Medium
#126086
opened Aug 25, 2026 by
VaggelisGian
Loading…
[XLA:GPU] Parallelize output dumping in FunctionalHloRunner.
#126085
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Refactor: Extract CalculateAccumulatorRegisters in GPU dot fusion cost model
#126084
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Fix SYSTEM pybind11 & flatbuffers
awaiting review
Pull request awaiting review
size:S
CL Change Size: Small
#126082
opened Aug 25, 2026 by
Flamefire
Contributor
Loading…
PR #46911: [XLA:GPU] Add GPU collectives FFI extension on the backend-agnostic API
#126081
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Integrate Triton up to a77e7c793abc0d0c923a9afb275058e2fe57a198
#126080
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
PR #47472: [oneDNN][TSL] Add AMX_FP8 CPU feature definition and CPUID detection
#126079
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
[XLA:CPU] Add a pass to remove size-1 dims in vectors.
#126076
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
[XLA:GPU] Run GpuAlgebraicSimplifier together with ZeroSizedHloElimination.
#126074
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Do not fuse power ops with users() > 1. This leads to duplicating the power op, which is expensive.
#126072
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Enforce minimum 32-bit hardware accumulator precision in GPU dot cost model
#126070
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Use two-pass variance calculation in BatchNormExpander to prevent negative variance with large input offsets.
#126066
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Fix configure.py crash when clang reports no parseable version
awaiting review
Pull request awaiting review
size:S
CL Change Size: Small
#126064
opened Aug 25, 2026 by
VaggelisGian
Loading…
Fix elu second derivative for tiny negative inputs
awaiting review
Pull request awaiting review
prtype:bugfix
PR to fix a bug
size:S
CL Change Size: Small
#126063
opened Aug 25, 2026 by
VaggelisGian
Loading…
[XLA] Add min_backend_config_size to HloProtoOptions and thread through MakeHloProto
#126062
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
PR #47794: Fix duplicate cupti_driver_cbid.h inclusion causing redefinition error
#126059
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Fix type confusion and use-after-free reference counting in pywrap_tfe Tape, VariableWatcher, and ForwardAccumulator bindings
#126057
opened Aug 25, 2026 by
copybara-service
Bot
•
Draft
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.