-
Notifications
You must be signed in to change notification settings - Fork 74.2k
Pull requests: tensorflow/tensorflow
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
#sdy Add a space between axes and device_ids to a mesh ASM print.
#75992
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Ensure TFL_CAPI_EXPORT is always placed at the start of the line.
#75991
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
[XLA:GPU] Enable latency hiding scheduler by default
#75990
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
[XLA:GPU] Minor readability improvements to double buffering loop unroller.
#75989
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Added quint8 and qint8 for mean_nd in xnnpack_delegate.
#75985
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Integrate LLVM at llvm/llvm-project@c59ac1a2f67d
#75984
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Integrate Triton up to [f4c48a92](https://github.com/openai/triton/commits/f4c48a9233957903e30474bae6443bf3d3a79bf7)
#75982
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Avoid duplicate work regarding Gemm autotuning.
#75981
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Avoid calling nccl NVTX_PAYLOAD_EVTATTR_SET externally
awaiting review
Pull request awaiting review
comp:xla
XLA
size:XS
CL Change Size: Extra Small
#75972
opened Sep 18, 2024 by
shawnwang18
Loading…
PR #17239: Remove erroneous bash invocation from LSP docs
#75970
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
[XLA:GPU] Increase the size limit for dot merger to infinity.
#75968
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Support dynamic and other kinds of broadcasting ops in fold_broadcast_to_pass
#75967
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Add Real-Time Optimization Feature to TensorFlow Grappler Optimizers
awaiting review
Pull request awaiting review
size:XL
CL Change Size:Extra Large
#75965
opened Sep 18, 2024 by
matthew-fitzgerald123
Loading…
Remove dependency on
-hlo-import-all-computations flag when exporting entry_computation_layout.
#75963
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Refactor
maybe_computation_propagation out of RunToFixPoint as MaybeComputationPropagation.
#75961
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
Replace any ;s that the unique name the Tf2Xla rewriter creates for new nodes.
#75960
opened Sep 18, 2024 by
copybara-service
bot
•
Draft
PR #16975: Add a few related optimization passes for fp8 gemm custom-calls.
#75959
opened Sep 17, 2024 by
copybara-service
bot
•
Draft
Augments StrategyGroup to model a one-to-many relationship between strategies and input shardings.
#75957
opened Sep 17, 2024 by
copybara-service
bot
•
Draft
Remove host offloading copy checks from HloVerifier. Due to float propagation and float normalization being separate passes, host offloading copies can be multi-precision (at least temporarily). Float normalization will fix that and insert a convert.
#75954
opened Sep 17, 2024 by
copybara-service
bot
•
Draft
Add a rule to algebraic simplifier to replace A - A with 0.
#75952
opened Sep 17, 2024 by
copybara-service
bot
•
Draft
Previous Next
ProTip!
What’s not been updated in a month: updated:<2024-08-18.