-
Notifications
You must be signed in to change notification settings - Fork 74k
Insights: tensorflow/tensorflow
Overview
Could not load contribution data
Please try again later
210 Pull requests merged by 2 people
-
Automated Code Change
#69008 merged
Jun 3, 2024 -
Implements the `DisassembleIntoSingleDeviceArrays` method for the BasicStringArray class.
#69016 merged
Jun 2, 2024 -
Fixes source list after schema utility functions relocation.
#68855 merged
Jun 2, 2024 -
[XLA:GPU] Clang-tidy cleanup for xla/service/algebraic_simplifier.cc
#68981 merged
Jun 1, 2024 -
[XLA:GPU] Clang-tidy cleanup for xla/service/gpu/command_buffer_scheduling.cc
#68980 merged
Jun 1, 2024 -
[XLA:GPU] Clang-tidy cleanup for xla/service/gpu/fusions/transpose.cc
#68979 merged
Jun 1, 2024 -
[XLA:GPU] Clang-tidy cleanup for xla/service/gpu/fusions/mlir/mlir_fusion_emitter.h
#68982 merged
Jun 1, 2024 -
[XLA:GPU] Clang-tidy cleanup for xla/service/gpu/fusions/in_place_dynamic_update_slice_mlir.h
#68978 merged
Jun 1, 2024 -
[XLA:GPU] Clang-tidy cleanup for xla/service/gpu/runtime/nccl_all_gather_thunk.cc
#68977 merged
Jun 1, 2024 -
Automated Code Change
#68935 merged
Jun 1, 2024 -
Automated Code Change
#68875 merged
Jun 1, 2024 -
[xla:cpu] Add TraceMe annotations to XLA:CPU thunks
#68916 merged
Jun 1, 2024 -
[xla:cpu] Add support for multi-output elemental loops
#68962 merged
Jun 1, 2024 -
Uses Stacker api to compute table stacking logic for SC
#68449 merged
May 31, 2024 -
PR #13244: [ROCm] Get rid of .bazelrc file in run_xla script
#68989 merged
May 31, 2024 -
Switch libtensorflow (Windows) to use Docker + Clang.
#68957 merged
May 31, 2024 -
Add more logging to remote tensor handle serialization
#68317 merged
May 31, 2024 -
[XLA:TPU] Fix a bug in GetGatherScatterBatchParallelDims.
#68748 merged
May 31, 2024 -
Fix a bug that tries to access RUNTIME_FUNCTION_REFS after deletion
#68974 merged
May 31, 2024 -
Support cast in Ifrt restore operation
#68905 merged
May 31, 2024 -
Allow optional KV store overwrites.
#68613 merged
May 31, 2024 -
[xla:cpu] Fix ir_emitter2 test crash
#68960 merged
May 31, 2024 -
#tf-data-sevice Downgrade null model warning to vlog.
#68963 merged
May 31, 2024 -
[TOCO Removal] Copy converter's Python API wrapper to a more appropriate dir.
#68841 merged
May 31, 2024 -
[xla:cpu] NFC: Remove deprecated linalg-based dot instruction codegen
#68924 merged
May 31, 2024 -
Remove unnecessary dependency accidentally included in a refactoring.
#68958 merged
May 31, 2024 -
Add max_enqueued_batches option for model servers.
#68577 merged
May 31, 2024 -
[XLA:GPU] Use xla_gpu.apply_indexing in ir_emitter_triton.
#68691 merged
May 31, 2024 -
BUILD file cleanup of old packaging rules.
#68918 merged
May 31, 2024 -
CSE after inlining.
#68944 merged
May 31, 2024 -
[XLA:GPU] Add ToString() methods to Tiled HLO instructions and computations.
#68800 merged
May 31, 2024 -
Remove compatibility constraint.
#68885 merged
May 31, 2024 -
[XLA:GPU] Use HloFusionAdaptor::GetParameters() instead of FindFusionArguments.
#68697 merged
May 31, 2024 -
Use DOxygen comments in lite/core/c/operator.h.
#68946 merged
May 31, 2024 -
Avoid rewrite ReadVaraibleOps that are assigned in the same module
#68929 merged
May 31, 2024 -
Roll forward without ConvertGenerator
#68953 merged
May 31, 2024 -
[MHLO] Add support for batching dims when legalizing `mhlo.gather` to linalg
#68869 merged
May 31, 2024 -
[XLA:GPU][NFC] Split logical RS if it is AR + DS to AR + DS + AR.
#68939 merged
May 31, 2024 -
Remove unused header
#68941 merged
May 31, 2024 -
Fix crash in unoptimized builds.
#68937 merged
May 31, 2024 -
Add runtime test for radix sort correctness
#68934 merged
May 31, 2024 -
Reverts adc512abf8c35c7c6874821c5fd6f6b26e371e1e
#68938 merged
May 31, 2024 -
Enable batched radix sort using CUB library
#68716 merged
May 31, 2024 -
Automated Code Change
#68644 merged
May 31, 2024 -
Remove obsolete TODO item.
#68932 merged
May 31, 2024 -
Remove obsolete TODO item.
#68931 merged
May 31, 2024 -
Automated Code Change
#68871 merged
May 31, 2024 -
Support DenseBoolArrayAttr in MLRT
#68906 merged
May 31, 2024 -
Add DynamicApproxTopKOp to StableHLO
#68865 merged
May 31, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68910 merged
May 31, 2024 -
[xla:cpu] Add support for a copy with layout change (traspose) to CopyThunk
#68919 merged
May 31, 2024 -
#tf-data-service Add a worker client test for alternative data transfer mechanisms.
#68912 merged
May 31, 2024 -
[xla:cpu] Add support for emitting reduction kernels
#68899 merged
May 31, 2024 -
[xla:gpu] Reuse CanTritonHandleGEMM in GpuAlgebraicSimplifier
#68915 merged
May 31, 2024 -
Automated g4 rollback of changelist 638510790.
#68922 merged
May 31, 2024 -
Update docstring to satisfy bulidifier
#68913 merged
May 30, 2024 -
[xla:pjrt:cpu] Correctly create PjRt memory for constant allocations
#68894 merged
May 30, 2024 -
Pattern-match FP8 calls with input converts without scales.
#68758 merged
May 30, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68886 merged
May 30, 2024 -
Add profiling_info_pb2 to lite.py
#68860 merged
May 30, 2024 -
First cl to implement a pass to unstack loop operands.
#68400 merged
May 30, 2024 -
Allow sharing the same IFRT client across multiple IFRT Proxy sessions
#68861 merged
May 30, 2024 -
Add note on TensorFlow NumPy 2.0 upgrade in the next release
#68851 merged
May 30, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68890 merged
May 30, 2024 -
Inline one-off use small functions as lambdas
#68489 merged
May 30, 2024 -
Reverts 73e2bb2a020aec9124c7f1c99a1eaaac6ab96a22
#68908 merged
May 30, 2024 -
Include stream_executor.h in a few files that were missing its inclusion.
#68850 merged
May 30, 2024 -
Reverts 8f8ce5cbf917285f77e1c08e75ffa011324565ef
#68904 merged
May 30, 2024 -
Support per-axis embedding lookup reference kernel
#68551 merged
May 30, 2024 -
Fix time_span filter in visibility filtering
#68901 merged
May 30, 2024 -
Integrate LLVM at llvm/llvm-project@765206e05045
#68853 merged
May 30, 2024 -
Fix test build.
#68843 merged
May 30, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68892 merged
May 30, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68891 merged
May 30, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68888 merged
May 30, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68893 merged
May 30, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68887 merged
May 30, 2024 -
[xla:cpu] Add support for emitting fusions with thread local computations
#68880 merged
May 30, 2024 -
[XLA:GPU] Fix logic for symbol lookup
#68876 merged
May 30, 2024 -
Make Stream a pure virtual base class to start eliminating circular dependencies between and StreamExecutor.
#68848 merged
May 30, 2024 -
Add flag to disable autotune for hlo compile tests.
#68879 merged
May 30, 2024 -
[xla:cpu] Add support for kWhile thunk
#68881 merged
May 30, 2024 -
[XLA] Simplify multihost_hlo_runner interface
#68708 merged
May 30, 2024 -
Update OpenXLA's Triton dependency to include the AMD backend.
#68877 merged
May 30, 2024 -
Automated Code Change
#68785 merged
May 30, 2024 -
[XLA] Propagate error properly via Status from MLIR pipeline
#68874 merged
May 30, 2024 -
[XLA] Update shmem memory usage requirements for new column reduction vector size
#68738 merged
May 30, 2024 -
[XLA:GPU] Add TritonSupportTest for dynamic-slice
#68720 merged
May 30, 2024 -
Slightly refactor IsSupportedF8Pattern in gemm_rewriter.cc.
#68571 merged
May 30, 2024 -
PR #11832: [XLA:CPU][oneDNN] Enable oneDNN MatMul + BiasAdd + Sigmoid fusion.
#68776 merged
May 30, 2024 -
[NFC] More informative error message on error from MLIR emitters
#68707 merged
May 30, 2024 -
Integrate StableHLO at openxla/stablehlo@c44d9af8
#68292 merged
May 30, 2024 -
[XLA:GPU] Add a "draft" version of a generic Triton Fusion Emitter.
#68799 merged
May 30, 2024 -
Automated Code Change
#68780 merged
May 30, 2024 -
Automated Code Change
#68794 merged
May 30, 2024 -
Reverts a03918a8b580ea180e16cd902f2c8f077b50e790
#68792 merged
May 30, 2024 -
[XLA:GPU] Add a GUnit matcher for TiledHloInstruction
#68823 merged
May 30, 2024 -
Disable MLIR-based TF graph optimizer
#68856 merged
May 30, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68812 merged
May 30, 2024 -
[TF] Remove unused TF->HLO conversion pattern
#68618 merged
May 30, 2024 -
Added a minor utility function to compute an elementwise product of two vectors.
#68774 merged
May 30, 2024 -
[stream_executor:host] Implement Launch method
#68540 merged
May 30, 2024 -
Implement `AssembleArrayFromSingleDeviceArrays` for BasicStringArray.
#68661 merged
May 30, 2024 -
Support paging in TfrtSavedModelServable.
#67061 merged
May 30, 2024 -
Allow model builder to build a model based on a file descriptor.
#68844 merged
May 30, 2024 -
[IFRT] Add several commonly used APIs to `xla::ifrt::Sharding`
#68845 merged
May 30, 2024 -
Change MatchShapeCoveringDynamicIndexInstruction to return the dynamic index if matched.
#68564 merged
May 30, 2024 -
Fix diff command for bazelrc in XLA build script
#68820 merged
May 30, 2024 -
PR #13065: [NVIDIA] Disable the bias reuse for fp8 dot
#68765 merged
May 30, 2024 -
PR #13107: [ROCm] Fix reduce_row_vectorized.hlo.test
#68849 merged
May 30, 2024 -
Update TFRT dependency to use revision
#68842 merged
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68778 merged
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68810 merged
May 29, 2024 -
Reverts changelist 638422085
#68846 merged
May 29, 2024 -
[XLA] Stop duplicating broadcasts before host memory offloading.
#68837 merged
May 29, 2024 -
[xla:cpu] Add support for kCall thunks
#68834 merged
May 29, 2024 -
[xla:cpu] Add support for kAbs and kReverse elemental host kernels
#68773 merged
May 29, 2024 -
[xla:cpu] Pack constants of sub-byte element type into dense storage compatible with XLA format
#68833 merged
May 29, 2024 -
[xla:gpu] Add support for u32 and u64 offset constants
#68840 merged
May 29, 2024 -
PR #13165: [ROCm] Update ROCm version to 6.0.2
#68821 merged
May 29, 2024 -
Load variable concurrently
#68839 merged
May 29, 2024 -
Delete duplicate schema.fbs (new location is //third_party/tensorflow/compiler/mlir/lite/schema/schema.fbs)
#68615 merged
May 29, 2024 -
Update mlir tests to reflect new namespace change
#67800 merged
May 29, 2024 -
move loop-invariant code outside the loop
#68835 merged
May 29, 2024 -
[xla:cpu] Add support for emitting host kernels for fusions
#68759 merged
May 29, 2024 -
[XLA:GPU] Do not CHECK in nvptx_compiler, we return Status already
#68789 merged
May 29, 2024 -
PR #13018: [ROCm] Distinguish between AMD and NVIDIA GPUs with relevant tags
#68815 merged
May 29, 2024 -
Remove Stream::platform_specific_stream method, and keep it on GpuStream where it makes sense.
#68813 merged
May 29, 2024 -
Use numpy as the source of truth for expm1 test.
#68819 merged
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68808 merged
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68811 merged
May 29, 2024 -
[IFRT] Add function for converting from xla::ifrt::Sharding to OpSharding.
#68822 merged
May 29, 2024 -
PR #13194: [ROCm] Fix build break in xla/service/gpu/ir_emitter_triton_rocm.cc
#68825 merged
May 29, 2024 -
Add simple loop peeling for imperfect tiling.
#68805 merged
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68809 merged
May 29, 2024 -
[xprof] Add infrastructure to render host offload transfers in trace viewer.
#68609 merged
May 29, 2024 -
[xla] Add `test_xla_cpu_thunks` tag to test XLA:CPU with thunks runtime
#68804 merged
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68807 merged
May 29, 2024 -
[xla:ffi] Renamed BufferBase to AnyBuffer
#68817 merged
May 29, 2024 -
Change IsCollectiveWithChannelId to return the instruction with channel_id or nullptr.
#68545 merged
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68767 merged
May 29, 2024 -
Create a simple loop unswitching pass.
#68795 merged
May 29, 2024 -
Sort summands by distance of symbols to insertion point.
#68796 merged
May 29, 2024 -
Combine sparsity patches
#68779 merged
May 29, 2024 -
Reduce logs in XNNPack delegate for prod builds.
#68735 merged
May 29, 2024 -
Define "bad indices handling" utilities.
#68594 merged
May 29, 2024 -
Revert usage of always_inline attribute.
#68793 merged
May 29, 2024 -
Automated Code Change
#68786 merged
May 29, 2024 -
Always inline functions that are called only once.
#68787 merged
May 29, 2024 -
More unsigned reduction fixes.
#68722 merged
May 29, 2024 -
Automated Code Change
#68637 merged
May 29, 2024 -
[JAX] Automatically share PGO data for GPU latency-hiding scheduler.
#64663 merged
May 29, 2024 -
[XLA:GPU] Pre-factoring: Add GetDnnVersionInfo helper that returns StatusOr.
#68746 merged
May 29, 2024 -
[XLA:GPU] Fix infinite loop in `ApproximateMatch` indexing test util.
#68783 merged
May 29, 2024 -
[XLA:GPU] Fix compute time calculation in the performance model.
#68590 merged
May 29, 2024 -
Automated Code Change
#68673 merged
May 29, 2024 -
[xla:cpu] Make HostKernel::Launch process all work in the main thread if thread_pool_ is nullptr.
#68750 merged
May 29, 2024 -
bad_indices_policy for GatherNd
#68576 merged
May 29, 2024 -
PR #13047: [XLA:GPU] Lowering cublasLt thunk to command buffer
#68731 merged
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68777 merged
May 29, 2024 -
Automated Code Change
#68676 merged
May 29, 2024 -
[xla:cpu] IrEmitter2: Get fast_min_max flag from hlo module config
#68747 merged
May 29, 2024 -
[xla:cpu] Add support for host kernels with multiple operands
#68761 merged
May 29, 2024 -
[xla:gpu] NFC: Simplify AddressComputationThunk by replacing llvm::zip with struct
#68730 merged
May 29, 2024 -
[xla:cpu] NFC: Give human readable names to LLVM IR instructions in host kernel prototype
#68760 merged
May 29, 2024 -
[xla:cpu] Add support for KernelThunk and jit-compiling elemental host kernels
#68664 merged
May 29, 2024 -
[xla:gpu] Fix static thread local leak in WhileThunk
#68766 merged
May 29, 2024 -
Add support for fusion-start and fusion-done operations in HLO.
#68763 merged
May 29, 2024 -
Remove fallback call to MLIR bridge.
#68751 merged
May 29, 2024 -
[xla:gpu] Detect dynamic slice offsets derived from loop iteration number and skip D2H transfer for them
#68733 merged
May 29, 2024 -
Fix checkfail in GatherV2
#65936 merged
May 29, 2024 -
Use opaque layout PJRT_Layouts_MemoryLayout in PjRtCApiBuffer::layout() to keep all the layout information.
#68681 merged
May 28, 2024 -
[pt composite lowerings]
#67805 merged
May 28, 2024 -
Check for int quant types before folding qconst into tpose conv.
#68756 merged
May 28, 2024 -
Remove StreamInterface::implementation method, and just cast based on the stream pointer.
#68568 merged
May 28, 2024 -
Re-enable eager test with TFRT:TPU
#68752 merged
May 28, 2024 -
Update robin_map dependency to v1.3.0.
#68570 merged
May 28, 2024 -
Integrate LLVM at llvm/llvm-project@9b79acedd689
#68734 merged
May 28, 2024 -
[XLA:GPU] Remove the `emit_param_load_fn` callback in `EmitTiledScope`.
#68724 merged
May 28, 2024 -
Break LegalizeTfWithTf2xla out of passes.h to make using it easier.
#68468 merged
May 28, 2024 -
Correct instrumentation of MLIR bridge ph1
#68289 merged
May 28, 2024 -
[xla:cpu] NFC: Merge HostKernelEmitter into IrEmitter2
#68628 merged
May 28, 2024 -
[XLA:GPU] Remove structured matchers for `SymbolicTile` tests and use string matcher instead.
#68744 merged
May 28, 2024 -
[XLA:GPU] Extract launch dimensions for SoftMax Triton fusions.
#68725 merged
May 28, 2024 -
[XLA:GPU] Extract tiled block creation in the Triton emitter into a standalone function and add tests.
#68723 merged
May 28, 2024 -
[xla:cpu] Add initial version of HostKernelEmitter to build host kernel LLVM IR
#68625 merged
May 28, 2024 -
[XLA:GPU] Remove unnecessary arg for the IndexingMap::Simplify.
#68635 merged
May 28, 2024 -
[xla] Annotate called computations with instruction type after flattening call graph
#68680 merged
May 28, 2024 -
Add post-partition optimization pass to TFRTSession.
#68607 merged
May 28, 2024 -
[Multi-host GPU]Integrate GPU topology into PjRtClient for multi-host GPU support
#67664 merged
May 28, 2024 -
[XLA:SPMD:CollectiveMatmul]
#68387 merged
May 28, 2024 -
NFC: BUILD file changes
#68736 merged
May 28, 2024 -
[xla:gpu] Do not copy constant dynamic-slice offsets from device memory
#68678 merged
May 28, 2024 -
[Triton] Clean-up duplicate bug reference
#68721 merged
May 28, 2024 -
[TSL] Do not cut the stack trace from Status when creating fatal error message
#68706 merged
May 28, 2024 -
[XLA:GPU] Add a util to reset programs count and deflake `gpu_compiler_test`.
#68728 merged
May 28, 2024 -
Add missing newline in the output of run_hlo_module
#68718 merged
May 28, 2024 -
Fix shuffle_reduce with unsigned arguments.
#68726 merged
May 28, 2024
159 Pull requests opened by 8 people
-
Add "bad_indcies_policy" to ScatterNd
#68682 opened
May 27, 2024 -
PR #13058: Use `BufferF32` instead of `Buffer` in GPU FFI example
#68683 opened
May 27, 2024 -
PR #12979: [NVIDIA] Fix PGLE for latency estimation of p2p instructions
#68684 opened
May 27, 2024 -
PR #13056: Use `operator->` with XLA FFI Result Buffers in custom call docs
#68685 opened
May 27, 2024 -
[XLA:GPU] Pass CUDA_VERSION explicitly into CudnnFusedConvRewriter.
#68687 opened
May 27, 2024 -
Reverts 3d88ed161be60dad3abc19c3f0da7790f780f1fc
#68689 opened
May 27, 2024 -
Don't fuse inside computations that are already fused.
#68692 opened
May 27, 2024 -
Reverts changelist 637594837
#68693 opened
May 27, 2024 -
PR #13089: Fix reduce_large_row_to_scalar.hlo.test
#68694 opened
May 27, 2024 -
PR #12940: [ROCm] Fix dot_bf16.hlo.test on ROCm
#68698 opened
May 27, 2024 -
[XLA:GPU] Pass CUDA and cuDNN versions (or ROCm counterparts) explicitly into CudnnFusedConvRewriter.
#68701 opened
May 27, 2024 -
Remove experimental_adaptive_avx_optimization flag from XNNPACK delegate options
#68703 opened
May 27, 2024 -
`ScatterNdFunctor` to update operands of all valid indices and continue on bad indices.
#68704 opened
May 27, 2024 -
Implement "bad_indcies_policy" for ScatterNd.
#68705 opened
May 27, 2024 -
Internal config change
#68712 opened
May 27, 2024 -
tf-lite CMake: add headers into sources to correctly install them
#68714 opened
May 28, 2024 -
Add support for TensorRT 10
#68715 opened
May 28, 2024 -
Remove experimental_adaptive_avx_optimization flag from XNNPACK delegate options
#68719 opened
May 28, 2024 -
Integrate LLVM at llvm/llvm-project@9b79acedd689
#68727 opened
May 28, 2024 -
[TEST] Remove xla_cc_test
#68729 opened
May 28, 2024 -
Do not block on bes upload success in ci jobs
#68737 opened
May 28, 2024 -
PR #13018: [ROCm] Distinguish between AMD and NVIDIA GPUs with relevant tags
#68739 opened
May 28, 2024 -
Clean up include and build file
#68740 opened
May 28, 2024 -
Automated Code Change
#68741 opened
May 28, 2024 -
Add xla_py_proto_library rule to xla.bzl.
#68743 opened
May 28, 2024 -
[XLA:CPU] Fix Test failure for oneDNN Convolutions with XLA
#68745 opened
May 28, 2024 -
Add utility for mapping from an unsigned integer type to its corresponding signed type.
#68754 opened
May 28, 2024 -
Add a function to get sparse core planes
#68755 opened
May 28, 2024 -
Add xla_py_proto_library rule to xla.bzl.
#68762 opened
May 28, 2024 -
Write data immediately in TFLite XNNPack delegate weight cache.
#68764 opened
May 29, 2024 -
Swap operands of dot if the LHS is fed by a parameter
#68768 opened
May 29, 2024 -
Automated Code Change
#68769 opened
May 29, 2024 -
Automated Code Change
#68770 opened
May 29, 2024 -
vtf-tf2-min-afs test test test
#68771 opened
May 29, 2024 -
Move XPlaneVisitor helper class to OSS.
#68772 opened
May 29, 2024 -
Automated Code Change
#68775 opened
May 29, 2024 -
Match floating-point comparer (checks for NaNs) in gpu_sort_rewriter pass
#68781 opened
May 29, 2024 -
Compile more kernels for radix sort on pairs (support u8 and f32 key types).
#68782 opened
May 29, 2024 -
Automated Code Change
#68784 opened
May 29, 2024 -
Automated Code Change
#68788 opened
May 29, 2024 -
Automated Code Change
#68791 opened
May 29, 2024 -
[XLA:GPU] Print Interval as a semi-open interval.
#68798 opened
May 29, 2024 -
Account for trivial dimensions in TransposesMinorDimension().
#68801 opened
May 29, 2024 -
OOB whatever; this CL description should be cleaned up before sent for review.
#68802 opened
May 29, 2024 -
[XLA:GPU] Use priority fusion in TritonGemmAutotunerExtractor.
#68803 opened
May 29, 2024 -
* Improve the perf of tf.sparse.segment_(mean/sum/sqrt) ops when using
#68814 opened
May 29, 2024 -
Stop using xla/status.h in favor of absl/status/status.h.
#68826 opened
May 29, 2024 -
Cap utilization at 100% when accurate bandwidth cannot be determined.
#68828 opened
May 29, 2024 -
copy the legalize tf passes under lite/stablehlo
#68829 opened
May 29, 2024 -
[XLA] Update fusion kind after DCE
#68830 opened
May 29, 2024 -
[XLA:MSA] Add a new function to estimate the total trip counts of instructions.
#68832 opened
May 29, 2024 -
Replace `tensorflow::Status` with `absl::Status` in `next_pluggable_device`
#68838 opened
May 29, 2024 -
[XLA] Limit DCE MoF cleanup to GPU backend
#68847 opened
May 29, 2024 -
PR #13065: [NVIDIA] Disable the bias reuse for fp8 dot
#68852 opened
May 30, 2024 -
[xla:gpu] Add GemvRewriter pass
#68854 opened
May 30, 2024 -
Automated Code Change
#68857 opened
May 30, 2024 -
Automated Code Change
#68858 opened
May 30, 2024 -
Automated Code Change
#68859 opened
May 30, 2024 -
Automated Code Change
#68863 opened
May 30, 2024 -
Automated Code Change
#68864 opened
May 30, 2024 -
Fixed description of num_threads in `tf.lite.interpreter` doc
#68866 opened
May 30, 2024 -
Integrate StableHLO at openxla/stablehlo@c44d9af8
#68868 opened
May 30, 2024 -
Automated Code Change
#68870 opened
May 30, 2024 -
PR #13109: Add flag for termination on nccl error
#68872 opened
May 30, 2024 -
Automated Code Change
#68873 opened
May 30, 2024 -
Automated Code Change
#68878 opened
May 30, 2024 -
[IFRT] Add an IFRT API for Topology.
#68884 opened
May 30, 2024 -
[XLA] Allow reusing channel IDs between sync/async variants of collectives.
#68889 opened
May 30, 2024 -
Add dependency for case_format test.
#68895 opened
May 30, 2024 -
PR #12952: [ROCM] hipblas-lt: bias pointer workaround
#68896 opened
May 30, 2024 -
#tf-data-service Add logic for dynamic SplitProviders.
#68898 opened
May 30, 2024 -
Use pthreadpool only when XNNPACK is enabled
#68900 opened
May 30, 2024 -
SparsifyModel returns absl::Status instead of TfLiteStatus.
#68902 opened
May 30, 2024 -
Improve error message when number of elements is not divisible by divisor.
#68903 opened
May 30, 2024 -
Integrate StableHLO at openxla/stablehlo@82bfdd48
#68907 opened
May 30, 2024 -
Integrate LLVM at llvm/llvm-project@b62ba7f5b1ca
#68911 opened
May 30, 2024 -
Move export_graphdef to tf2xla directory.
#68914 opened
May 30, 2024 -
Revert "Use MLIR op lowering of FusedBatchNormV3 and FusedBatchNormGradV3 for tf2xla"
#68920 opened
May 30, 2024 -
Reverts adc512abf8c35c7c6874821c5fd6f6b26e371e1e
#68921 opened
May 30, 2024 -
Apply symmetric check for hybrid embedding_lookup only
#68925 opened
May 31, 2024 -
[XLA:GPU] Emit matrix-vector multiplication as GemmFusion
#68926 opened
May 31, 2024 -
Reverts bf42ffa578b487183e7bf0635784dc0ed3adac09
#68927 opened
May 31, 2024 -
an attempt to change the doc for "bad_indices_policy"
#68928 opened
May 31, 2024 -
Automated Code Change
#68930 opened
May 31, 2024 -
[exp] align the behavior on CPU and GPU for scatternd, gathernd
#68933 opened
May 31, 2024 -
[XLA:GPU] Add initial support for constraints in `SymbolicTile`.
#68936 opened
May 31, 2024 -
Automated Code Change
#68940 opened
May 31, 2024 -
[XLA:GPU] Simplify error handling for GetSymbol, remove redundant logging
#68942 opened
May 31, 2024 -
[XLA:GPU] Use `MatchReduceScatter` primitive in all-reduce-splitter pass.
#68943 opened
May 31, 2024 -
Automated Code Change
#68945 opened
May 31, 2024 -
Automated Code Change
#68947 opened
May 31, 2024 -
replace the invalid indices with oob values for spars_grouping_packed
#68948 opened
May 31, 2024 -
PR #13258: Give GPU compiler class access to PJRT key-value store.
#68949 opened
May 31, 2024 -
Integrate LLVM at llvm/llvm-project@ae8627809076
#68951 opened
May 31, 2024 -
Change norm sharding for llama2-7b to fsdp.
#68952 opened
May 31, 2024 -
[XLA:GPU] Reenable dot_algorithm_support_test in OSS
#68954 opened
May 31, 2024 -
[XLA:GPU] Remove dead code
#68955 opened
May 31, 2024 -
Use get_api_version to choose gather/scatter
#68956 opened
May 31, 2024 -
Rename StreamExecutorInterface to be StreamExecutor to save on all the renaming elsewhere.
#68961 opened
May 31, 2024 -
[XLA] Rollback simplification of multihost_hlo_runner interface until callers are migrated.
#68964 opened
May 31, 2024 -
Integrate LLVM at llvm/llvm-project@ae8627809076
#68965 opened
May 31, 2024 -
Revert bytecode changes from #2259
#68967 opened
May 31, 2024 -
PR #11583: Weight offloading of Jax memories: support memory_kind for GPUs
#68968 opened
May 31, 2024 -
[xla:cpu] Add support for Infeed thunk to xla:cpu runtime
#68969 opened
May 31, 2024 -
[PJRT C API] Move eligible Device tests from tpu to the test factory
#68970 opened
May 31, 2024 -
Reimplementation of ICI weight distribution.
#68971 opened
May 31, 2024 -
Implement a batched version of `Array::GetReadyFuture()`
#68972 opened
May 31, 2024 -
[XLA:GPU] Fix hlo_fusion_analysis_test.
#68975 opened
May 31, 2024 -
[XLA:GPU] Clang-tidy cleanup for xla/service/gpu/cudnn_fusion_compiler.{cc,h}
#68976 opened
May 31, 2024 -
[xla:cpu] Add support for emitting kMap and kSlice and enable more tests
#68983 opened
May 31, 2024 -
Bump requests from 2.31.0 to 2.32.2
#68984 opened
May 31, 2024 -
Remove key_value_set overload for byte values.
#68985 opened
May 31, 2024 -
[JAX] Fail gracefully when an array with multiple shards is passed to make_array_from_single_device_arrays.
#68986 opened
May 31, 2024 -
Update GPU compatibility checker for CAST op
#68987 opened
May 31, 2024 -
[jax:cpu] Add a test config to run JAX with new XLA:CPU runtime
#68988 opened
May 31, 2024 -
PR #13293: [ROCm] Add gpu_plugin dependency for AMD gpus.
#68990 opened
May 31, 2024 -
Tighten error tolerances for various functions, especially on TPUs starting with TPUv6.
#68991 opened
May 31, 2024 -
Move tf_to_tfl_flatbuffer.cc from TfLiteStatus to absl::Status
#68994 opened
Jun 1, 2024 -
Support int64 indice type for tflite dynamic update slice op.
#68995 opened
Jun 1, 2024 -
Move `::mlir::lite::QuantizeWeights` from `TfLiteStatus` to `absl::Status`.
#68996 opened
Jun 1, 2024 -
[XLA:GPU] Renaming AddressComputationThunk to DynamicSliceThunk
#68997 opened
Jun 1, 2024 -
Automated Code Change
#68998 opened
Jun 1, 2024 -
Automated Code Change
#68999 opened
Jun 1, 2024 -
Automated Code Change
#69000 opened
Jun 1, 2024 -
Automated Code Change
#69001 opened
Jun 1, 2024 -
Automated Code Change
#69002 opened
Jun 1, 2024 -
Automated Code Change
#69003 opened
Jun 1, 2024 -
Automated Code Change
#69004 opened
Jun 1, 2024 -
Automated Code Change
#69005 opened
Jun 1, 2024 -
Automated Code Change
#69006 opened
Jun 1, 2024 -
Automated Code Change
#69007 opened
Jun 1, 2024 -
Automated Code Change
#69009 opened
Jun 1, 2024 -
Automated Code Change
#69010 opened
Jun 1, 2024 -
Support passing user-supplied data to TfLiteOperator's methods
#69011 opened
Jun 1, 2024 -
Automated Code Change
#69012 opened
Jun 1, 2024 -
Automated Code Change
#69013 opened
Jun 1, 2024 -
Fix a bug in AtomicFunction destructor.
#69014 opened
Jun 1, 2024 -
PR #13302: [GPU] Fix scoped timer of LLVM verifier and move it to a higher logging level.
#69015 opened
Jun 1, 2024 -
Automated Code Change
#69017 opened
Jun 2, 2024 -
Automated Code Change
#69018 opened
Jun 2, 2024 -
Automated Code Change
#69019 opened
Jun 2, 2024 -
Automated Code Change
#69023 opened
Jun 2, 2024 -
Automated Code Change
#69024 opened
Jun 2, 2024 -
Automated Code Change
#69025 opened
Jun 2, 2024 -
Automated Code Change
#69026 opened
Jun 2, 2024 -
Automated Code Change
#69027 opened
Jun 2, 2024 -
Automated Code Change
#69028 opened
Jun 2, 2024 -
Automated Code Change
#69029 opened
Jun 2, 2024 -
Automated Code Change
#69030 opened
Jun 2, 2024 -
[xla:ffi] Add filegroup for all FFI headers
#69031 opened
Jun 3, 2024 -
Automated Code Change
#69032 opened
Jun 3, 2024
42 Issues closed by 17 people
-
dynamic input shape with InferenceRunner
#68695 closed
Jun 2, 2024 -
[xrt] Add "Free all memory" op
#23562 closed
Jun 1, 2024 -
[DeepLabV3+] How to get class probabilities in C++ using frozen graph?
#23814 closed
Jun 1, 2024 -
Secondary algorithm not provided error when not using cuDNN autotune
#24044 closed
Jun 1, 2024 -
TextLineDataset supporting start/end file positions
#24198 closed
Jun 1, 2024 -
Multiple CNN models concurrently on the GPU
#66998 closed
Jun 1, 2024 -
No kernel identified while running tensorflow with GPU
#67768 closed
Jun 1, 2024 -
No equivalent to theano.tensor.slinalg.Eigvalsh
#18083 closed
May 31, 2024 -
tf.keras.backend.eye crash (abort) with large input
#57711 closed
May 31, 2024 -
Unable bazel build tensorflow with cuda support inside a docker
#67145 closed
May 31, 2024 -
Broken ExtensionType interoperability with Keras in 2.16
#67334 closed
May 31, 2024 -
tf.keras.utils.Sequence or tf.keras.utils.PyDataset is abnormal in tensorflow2.16
#67386 closed
May 31, 2024 -
Building binary for Sparkfun Edge
#67418 closed
May 31, 2024 -
Does anyone know which version of ROCM is used to build the tensorflow-rocm ?
#67620 closed
May 31, 2024 -
Remove resizeInput interface to TFLite Interpreter
#62303 closed
May 30, 2024 -
TPU v3-8 CrossReplicaSum_33 Error
#68210 closed
May 30, 2024 -
Bazel always rebuild all code even I just modify single cpp file when config=cuda is used
#26498 closed
May 30, 2024 -
[RNN] LSTM Model conversion error after upgrading to tf 2.16.1 from 2.15
#68797 closed
May 30, 2024 -
RuntimeError when invoking TFLite INT8 model with tile operation
#68474 closed
May 30, 2024 -
problem with converted custom model at TensorFlow Lite
#67086 closed
May 30, 2024 -
Fit function to work with adaptive step size
#67252 closed
May 30, 2024 -
Bazel - Tensorflow build failed
#67554 closed
May 30, 2024 -
Tensorflow using Intel GPU instead of NVIDIA GPU
#67558 closed
May 30, 2024 -
Errors running `bazel run tensorflow/examples/speech_commands:train`
#63572 closed
May 29, 2024 -
Issue with Tesnorflow JS Face Detection on Production
#68717 closed
May 29, 2024 -
_image_ops.so not found
#67136 closed
May 29, 2024 -
Why don't you provide a standard go-lang get install?
#67221 closed
May 29, 2024 -
Documentation for `tf.linalg.set_diag()` is missing return information
#67255 closed
May 29, 2024 -
L0i
#67426 closed
May 29, 2024 -
`Check failed` in `tf.raw_ops.GatherV2` when the value of `axis` is too large.
#65869 closed
May 29, 2024 -
"CUDA_ERROR_NOT_FOUND: named symbol not found" in Docker container
#68711 closed
May 28, 2024 -
Proposal: Making the cmake build distribution friendly
#13061 closed
May 28, 2024 -
Aborted (core dumped) in `tf.irfftnd/ifftnd`
#67098 closed
May 28, 2024 -
Aborted (core dumped) in `tf.raw_ops.IRFFTND\RFFTND\FFTND\IFFTND`
#68648 closed
May 28, 2024 -
Aborted (core dumped) in `tf.fftnd/rfftnd`
#68647 closed
May 28, 2024 -
Tensorflow flex so file build for arm64 target
#68258 closed
May 28, 2024 -
libhexagon_interface.so for non Android - eLinux platform
#66872 closed
May 28, 2024 -
DeprecationWarning from google.protobuf
#68194 closed
May 27, 2024 -
I was using tf lite and trying to get the examples but cant
#68585 closed
May 27, 2024 -
wrong calculations with TPU distribution strategy
#67301 closed
May 27, 2024
17 Issues opened by 16 people
-
ERROR CONVERTING SYMBOLIC TENSOR TO NUMPY ARRAY
#69022 opened
Jun 2, 2024 -
Running an Integrated Image Segmenter in Java
#69021 opened
Jun 2, 2024 -
tflite-runtime 2.11 python wheel for windows
#69020 opened
Jun 2, 2024 -
Current tensorflow[and-cuda] installed by pip pulls ptxas which causes Jupyter kernel restart
#68966 opened
May 31, 2024 -
TypeError: Expected int32, got 1e-07 of type 'float' instead.
#68959 opened
May 31, 2024 -
__add__ with floating point values
#68923 opened
May 30, 2024 -
Not able to build TensorFlow with GPU support
#68867 opened
May 30, 2024 -
Wrong explanation about an argument of tflite interpreter
#68862 opened
May 30, 2024 -
segmentation fault when tf.histogram_fixed_width receives large `value_range` and `nbins` on CPU mode
#68836 opened
May 29, 2024 -
Bazel compiling source code failed because of highwayhash/sip_hash.cc
#68824 opened
May 29, 2024 -
Training model with the Poisson loss function and the Adam optimizer resulted in NaN loss
#68806 opened
May 29, 2024 -
TensorFlow Cuda in Docker under WSL2 not wokring
#68710 opened
May 27, 2024 -
There is no target called wheel
#68702 opened
May 27, 2024 -
Trouble Running TensorFlow v2.16.1 with NVIDIA GeForce 940MX GPU #914
#68696 opened
May 27, 2024
194 Unresolved conversations
Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.
-
remove flatc-generated files from source
#68665 commented on
May 30, 2024 • 7 new comments -
Can't install tflite-model-maker from pip or github
#59855 commented on
Jun 1, 2024 • 5 new comments -
NumPy 2.0 support
#67291 commented on
May 31, 2024 • 4 new comments -
Wrong quantized_dimension (axis) when "per-channel" quantization
#66081 commented on
May 29, 2024 • 4 new comments -
TF 2.16.1 Fails to work with GPUs
#63362 commented on
May 31, 2024 • 4 new comments -
Unable to Force-load TensorFlowLiteSelectTfOps.framework, created with Selective Build, in iOS
#67790 commented on
May 31, 2024 • 4 new comments -
Couldn't resolve TF-TRT Warning: Could not find TensorRT
#68335 commented on
May 30, 2024 • 4 new comments -
tflite-model-maker cannot be installed correctly for different reasons on different configurations
#61719 commented on
May 30, 2024 • 4 new comments -
Error Handling suggestion
#68606 commented on
May 29, 2024 • 3 new comments -
Remove null pointer from error message in mlir::odml
#67556 commented on
May 29, 2024 • 3 new comments -
TFLiteConverter produces model that doesn't conform to GPUv2 (TfLiteGpuDelegate Init: FULLY_CONNECTED: Amount of input channels should match weights width)
#66729 commented on
May 31, 2024 • 3 new comments -
Local rendezvous is aborting with status: OUT_OF_RANGE: End of sequence when call some methods of `tf.data`
#68593 commented on
May 31, 2024 • 3 new comments -
Inference time using Interpreter API on Android inconsistent and 10–50 times slower than same tflite model on iOS
#66025 commented on
May 30, 2024 • 3 new comments -
Tensorflow lite in Android App
#67699 commented on
May 29, 2024 • 2 new comments -
Memory leak in custom training loop + tf.function
#50765 commented on
May 29, 2024 • 2 new comments -
Any build help with tensorflow-text
#67465 commented on
May 29, 2024 • 2 new comments -
Fix a bug in AsyncAllocator memory pools access setting.
#67373 commented on
May 28, 2024 • 2 new comments -
Why does my GPU trained model on Jetson Nano preform so much worse than CPU.
#68610 commented on
May 30, 2024 • 2 new comments -
TensorFlowLiteSelectTfOps - Compile error
#67748 commented on
May 31, 2024 • 2 new comments -
GPUv2 segfaults on split-head attention CLIP model
#66721 commented on
May 29, 2024 • 2 new comments -
TFLite Op type not registered (RegexSplitWithOffsets) in Swift
#65475 commented on
May 29, 2024 • 2 new comments -
TFLite for LSTM: Downscale accumulation from 32-bit to 16-bit before applying to activation
#68670 commented on
May 27, 2024 • 2 new comments -
TensorFlow 2.15 cannot be imported after Poetry install on Windows
#62899 commented on
Jun 1, 2024 • 2 new comments -
Fail to run person_detect.tflite via TFLite benchmark_model
#68308 commented on
May 27, 2024 • 2 new comments -
Tensorflow lite GPU delegate can't work well and NNAPI can't make the model faster
#66702 commented on
May 31, 2024 • 2 new comments -
output_padding argument in Conv1DTranspose
#68505 commented on
May 30, 2024 • 2 new comments -
tensorflow-macos still required for version 2.16.1
#63495 commented on
May 30, 2024 • 2 new comments -
Tensorflow Docker Instructions
#68420 commented on
May 28, 2024 • 2 new comments -
Aborted (core dumped) in tf.io.TFRecordWriter
#63337 commented on
May 30, 2024 • 2 new comments -
Block Matrix / Block Diagonal Matrix
#24364 commented on
May 31, 2024 • 1 new comment -
tf.raw_ops.SparseConcat: Overflow bug
#63377 commented on
May 31, 2024 • 1 new comment -
Tensorflow C API: SessionRun batch size (how to properly set)
#22869 commented on
May 31, 2024 • 1 new comment -
Error with transform_graph tool: "Failed to parse --transform argument"
#22843 commented on
May 31, 2024 • 1 new comment -
crash via tf_should_use format_stack
#22770 commented on
May 31, 2024 • 1 new comment -
Unexpected behavior in tf.data.Dataset.from_generator when there is a naming collision
#67566 commented on
May 31, 2024 • 1 new comment -
Empty tfrecord in gcs would fail the TFRecordDataset.read
#22598 commented on
May 31, 2024 • 1 new comment -
Consider automatic casting rules for promoting dtypes
#21491 commented on
May 31, 2024 • 1 new comment -
How can I clear GPU memory in tensorflow 2?
#36465 commented on
May 31, 2024 • 1 new comment -
Docs on using TPUs with custom training loop can be misleading
#42144 commented on
May 31, 2024 • 1 new comment -
TypeError: Can't convert Operation 'MutableHashTable' to Tensor
#24439 commented on
May 31, 2024 • 1 new comment -
Improve tf.data graph representation in TensorBoard
#21349 commented on
May 31, 2024 • 1 new comment -
Allow full deallocation of GPU memory
#15880 commented on
May 31, 2024 • 1 new comment -
TFlite minimal example failing on latest tensorflow repo
#62769 commented on
May 31, 2024 • 1 new comment -
Feature request: segment_argmax
#13460 commented on
May 31, 2024 • 1 new comment -
Dataset Unzip Operation
#12851 commented on
May 31, 2024 • 1 new comment -
[feature request] make MutableHashTableOfScalars and MutableHashTableOfTensors trainable
#24539 commented on
May 31, 2024 • 1 new comment -
Unable to compile a quantized graph using XLA AOT?
#11604 commented on
May 31, 2024 • 1 new comment -
Feature Request: Add separable_conv2d_transpose operation
#12001 commented on
May 31, 2024 • 1 new comment -
PEP 484 Type Annotations (feature request)
#12345 commented on
May 31, 2024 • 1 new comment -
The changes made are as follows:
#59733 commented on
May 29, 2024 • 1 new comment -
[tflite] Add png to tflite evaluation tool
#60193 commented on
May 29, 2024 • 1 new comment -
add missing floor operation
#60317 commented on
May 29, 2024 • 1 new comment -
[TOSA] bug fix convert-tfl-uint8 and TFL Cast legalization
#60701 commented on
May 29, 2024 • 1 new comment -
[TOSA] Matching ranks for TOSA operators
#60753 commented on
May 29, 2024 • 1 new comment -
[TfLite][GPU/CPU] Added GroupNormalization CPU/GPU kernel
#60794 commented on
May 29, 2024 • 1 new comment -
[MHLO] let unfuse-batch-norm-inference produce bcast_mul+bcast_add
#61120 commented on
May 29, 2024 • 1 new comment -
Add ComplexOp as a builtin to TFLite
#61359 commented on
May 29, 2024 • 1 new comment -
[TFLite] Fix Android build with CMake
#61381 commented on
May 29, 2024 • 1 new comment -
Support for arbitrary tensor sizes in `tfl.strided_slice` lowering
#61424 commented on
May 29, 2024 • 1 new comment -
[TOSA] Fix tfl.transpose_conv legalization when layer has bias
#61493 commented on
May 29, 2024 • 1 new comment -
Use `tf.Unpack` when unrolling `tf.BatchMatMul`
#61497 commented on
May 29, 2024 • 1 new comment -
Add RISC-V platform support with cmake in TFLite
#61581 commented on
May 29, 2024 • 1 new comment -
Add `use_default_shell_env = True` to `ctx.actions.run`
#61591 commented on
May 29, 2024 • 1 new comment -
Add support for quantization when real_output_multiplier >= 1 in lite/kernels/add.cc
#61698 commented on
May 29, 2024 • 1 new comment -
lite: add tensorflowlite_flex to minimal example
#62484 commented on
May 29, 2024 • 1 new comment -
Tensor pull
#68316 commented on
Jun 1, 2024 • 1 new comment -
TFLite GPUv2: ADD(x, 1e-5) results in severely wrong output
#67216 commented on
May 31, 2024 • 1 new comment -
aarch64/arm64 Tensorflow Lite runtime wheels not compatible with silicon OSX ?
#67422 commented on
Jun 1, 2024 • 1 new comment -
IllegalArgumentException: Internal error: Failed to run on the given Interpreter
#66594 commented on
Jun 1, 2024 • 1 new comment -
axis argument for FFT ops (tf.signal.fft, tf.signal.fft2d, etc.)
#23156 commented on
Jun 1, 2024 • 1 new comment -
Intermittent very long latency in XRT operations
#22975 commented on
Jun 1, 2024 • 1 new comment -
Can't use CTCBeamSearchDecoder in c++, LINK ERROR occur,BUG in CTCBeamSearchDecoder 's source code
#22894 commented on
Jun 1, 2024 • 1 new comment -
Potential use-after-free bug in function `EagerTensor_shape_tuple` by using an object after all its references released (a static analyzer report)
#52085 commented on
Jun 1, 2024 • 1 new comment -
Multi-arch docker images
#14934 commented on
Jun 2, 2024 • 1 new comment -
[TFLite] Add int8 and int16x8 support for the ROUND operator
#50786 commented on
May 29, 2024 • 1 new comment -
[TFLite] Add int16x8 support for RELU_N1_TO_1 and PRELU operators
#52913 commented on
May 29, 2024 • 1 new comment -
[TFLite] Add int16x8 support for DEPTH_TO_SPACE and SPACE_TO_DEPTH operators
#53011 commented on
May 29, 2024 • 1 new comment -
Add lowering for tf.ImageProjectiveTransformV3 for translation
#55335 commented on
May 29, 2024 • 1 new comment -
Allow non-empty stderr when locating numpy include path (closes #56373)
#56347 commented on
May 29, 2024 • 1 new comment -
Add int16x8 support for GELU operator
#58294 commented on
May 29, 2024 • 1 new comment -
stateful random ops, fix crash for empty output
#58336 commented on
May 29, 2024 • 1 new comment -
[oneDNN] [tfg] [remapper] Add Conv2D + BiasAdd fusion in MLIR
#59528 commented on
May 29, 2024 • 1 new comment -
Use `tf` libaries instead of `array_ops` to provide compatibility with the Tensorflow API
#59638 commented on
May 29, 2024 • 1 new comment -
TFLite Multipose model input error for android
#66564 commented on
May 31, 2024 • 1 new comment -
GPUv2 syntax error in shader generation in broadcast with different batch sizes
#66712 commented on
May 29, 2024 • 1 new comment -
Op support request: Matmul with constant left hand side
#66727 commented on
May 29, 2024 • 1 new comment -
preprocessor definition clash with glog
#7480 commented on
May 29, 2024 • 1 new comment -
Feature Request: use S3 for checkpoint loading/saving
#13796 commented on
May 29, 2024 • 1 new comment -
Feature Request: C++ gradient for Cast
#14081 commented on
May 29, 2024 • 1 new comment -
[Fetaure Request] Bicubic interpolation using map coordinates
#14482 commented on
May 29, 2024 • 1 new comment -
gRPC debug URL scheme support for Windows
#17933 commented on
May 29, 2024 • 1 new comment -
Feature Request: Support for configuring deterministic options of cudNN Conv routines
#18096 commented on
May 29, 2024 • 1 new comment -
ValueError: as_list() is not defined on an unknown TensorShape. during training
#68217 commented on
May 27, 2024 • 1 new comment -
Request a new padding mode
#18213 commented on
May 29, 2024 • 1 new comment -
[Feature request] unsquashing unsorted_segment_x
#18311 commented on
May 29, 2024 • 1 new comment -
Riccati solver suggest
#23735 commented on
May 29, 2024 • 1 new comment -
import tensorflow.compat.v2 could not be resolved
#67331 commented on
May 29, 2024 • 1 new comment -
Linker error when tensorflow is being built with clang-17 inside anaconda environment
#68246 commented on
May 29, 2024 • 1 new comment -
Apple Silicon, building pip package ... clang: error: linker command failed with exit code 1
#67473 commented on
May 28, 2024 • 1 new comment -
Segmentation fault when using tflite_model_maker searcher.TextDataLoader.create(EmbeddingModel, l2_normalize=True)
#65409 commented on
May 27, 2024 • 1 new comment -
how to solve this error ? : AttributeError: module 'keras._tf_keras.keras.layers' has no attribute 'experimental'
#64349 commented on
May 28, 2024 • 1 new comment -
Running the same model in TF and TFLiteMicro produces different outputs
#66358 commented on
May 27, 2024 • 1 new comment -
tf.keras cannot weight classes when using multiple outputs
#41448 commented on
May 27, 2024 • 1 new comment -
Add support for TensorRT 10
#66473 commented on
May 28, 2024 • 1 new comment -
Unable to install old version of tensorflow
#66950 commented on
May 27, 2024 • 1 new comment -
can't crosscompile tensorflow lite c library using camke
#68616 commented on
May 27, 2024 • 1 new comment -
Android Tflite model fails to load on GPU Delegate: CL_OUT_OF_HOST_MEMORY
#68470 commented on
May 27, 2024 • 1 new comment -
make error in Tensorflow Lite Micro
#66616 commented on
May 28, 2024 • 1 new comment -
Getting [java.lang.IllegalArgumentException: Internal error: Error applying delegate] error while applying GPU Delegate
#65648 commented on
May 28, 2024 • 1 new comment -
How to reduce the time running invoke()
#68424 commented on
May 28, 2024 • 1 new comment -
TFLite ConvTranspose3D implemented typo
#68319 commented on
May 28, 2024 • 1 new comment -
RuntimeError when invoking TFLite INT8 model with tile operation
#67789 commented on
May 28, 2024 • 1 new comment -
No gradient defined for operation 'MatrixExponential' (op type: MatrixExponential)
#15465 commented on
May 27, 2024 • 1 new comment -
Tensorflow Developer certificate didnt recieved yet
#68654 commented on
May 28, 2024 • 1 new comment -
404 - page not found
#66708 commented on
May 31, 2024 • 1 new comment -
[tensorflow/tflite branch r2.10] crash at absl::OkStatus();
#66990 commented on
May 31, 2024 • 1 new comment -
Improve tf_compile to allow emitting HLO module instead of/in addition to executable
#26627 commented on
May 30, 2024 • 1 new comment -
tf.train.BytesList should accept bytearray as an input type
#27047 commented on
May 30, 2024 • 1 new comment -
tensorflow[and-cuda] 2.15.0/2.15.1 compatibility with jax[cuda12]
#68290 commented on
May 30, 2024 • 1 new comment -
[feature] Smarter Handling of Image Data Format
#8227 commented on
May 27, 2024 • 1 new comment -
einsum with ellipsis is slow
#48484 commented on
May 30, 2024 • 1 new comment -
Include C and C++ APIs with binary distributions
#18360 commented on
May 30, 2024 • 1 new comment -
Slot variables used in an optimizer must have the same shape with the variable to be optimized?
#19457 commented on
May 30, 2024 • 1 new comment -
set_intersection doesn't work as expectation - tensorflow 1.6.0
#19832 commented on
May 30, 2024 • 1 new comment -
[BUG] tf.train.Saver in non-local filesystem
#20043 commented on
May 30, 2024 • 1 new comment -
Feature Request: Document Best Practice For Feeding New Data to a Restored Metagraph
#20098 commented on
May 30, 2024 • 1 new comment -
Use shallow clones from git repos in CMake build
#20222 commented on
May 30, 2024 • 1 new comment -
fused_batch_norm's gradient implementation is incomplete
#20407 commented on
May 30, 2024 • 1 new comment -
TimeDistributed layer with nested model no longer working in TensorFlow 2.16.1
#65393 commented on
May 29, 2024 • 1 new comment -
Instructions to install Tensorflow
#68394 commented on
May 29, 2024 • 1 new comment -
Documentation for `tf.dynamic_partition()` has examples without a call to the API being described
#68259 commented on
May 29, 2024 • 1 new comment -
Support for 16 bit activations in ExpandDims operation.
#68293 commented on
May 29, 2024 • 1 new comment -
embed_sequence and embedding_lookup behave differently on CPU vs. GPU
#17417 commented on
May 27, 2024 • 1 new comment -
TypeError: this __dict__ descriptor does not support '_DictWrapper' objects during trivial model save
#63548 commented on
May 30, 2024 • 1 new comment -
Can not find strtod_l function on Android device
#61951 commented on
May 30, 2024 • 1 new comment -
Segmentation fault (core dumped) in `tf.raw_ops.SoftmaxCrossEntropyWithLogits`
#67531 commented on
May 30, 2024 • 1 new comment -
TensorFlow 2.16 is not found in Maven Central
#65694 commented on
May 30, 2024 • 1 new comment -
android.support.v13.app does not exist while building TfLiteCameraDemo
#64019 commented on
May 30, 2024 • 1 new comment -
Saved model won't load: Unable to synchronously open object (bad local heap signature)
#63365 commented on
May 30, 2024 • 1 new comment -
tf.shape output is wrong when net input shape is changed during import
#21185 commented on
May 30, 2024 • 1 new comment -
Remove waiting for the remote address of a `TensorHandle` from within the scope of acquiring a shared lock in `RemoteMgr`
#67703 commented on
May 31, 2024 • 0 new comments -
Disable some long running tests that would otherwise timeout
#68307 commented on
May 31, 2024 • 0 new comments -
Modify DirectPluginOpKernelContext constructor to get more type safety
#67280 commented on
May 31, 2024 • 0 new comments -
Switch libtensorflow (Windows) to use Docker + Clang.
#67845 commented on
May 31, 2024 • 0 new comments -
Support Python 3.12
#62003 commented on
May 30, 2024 • 0 new comments -
Extract utility classes from TraceMeRecorder.
#67856 commented on
May 31, 2024 • 0 new comments -
Run OpenGL-based inference always on the current MP GL context/thread.
#67562 commented on
May 31, 2024 • 0 new comments -
[xla:gpu] Track loop iteration counter of a WhileThunk in thread local variable
#68679 commented on
May 27, 2024 • 0 new comments -
Automated Code Change
#68675 commented on
May 29, 2024 • 0 new comments -
[XLA:GPU][IndexAnalysis] Use a flag for IsKnownEmpty instead of recomputing every time.
#68669 commented on
May 28, 2024 • 0 new comments -
Automated Code Change
#68656 commented on
May 29, 2024 • 0 new comments -
[XLA:GPU] Dead code cleanup for xla/service/gpu/runtime/command_buffer_cmd
#68622 commented on
May 28, 2024 • 0 new comments -
[xla:cpu] Add initial support for Thunk-based execution to CpuCompiler and CpuExecutable
#68611 commented on
May 27, 2024 • 0 new comments -
Save function argument names as locs through MHLO<->HLO conversion.
#68605 commented on
May 31, 2024 • 0 new comments -
Implement support for "ignore bad indices" case for GatherNd.
#68596 commented on
May 31, 2024 • 0 new comments -
Remove usage of --xla_gpu_enable_triton_hopper in autotuner
#68581 commented on
May 27, 2024 • 0 new comments -
#tf-data-sevice Downgrade null model warning to vlog.
#68539 commented on
May 31, 2024 • 0 new comments -
Add a device_layout parameter to BufferFromHostLiteral in pjirt_client.
#68476 commented on
May 29, 2024 • 0 new comments -
When serializing a remote tensor handle, stop passing device name explicitly where it can be inferred
#68472 commented on
May 31, 2024 • 0 new comments -
Use rocm_library in rocm_embedded_test_modules
#68408 commented on
May 28, 2024 • 0 new comments -
Move XPlaneVisitor helper class to OSS.
#68405 commented on
May 29, 2024 • 0 new comments -
Add new build script to XLA
#68393 commented on
May 31, 2024 • 0 new comments -
Bump fullyconnect operation version to 12.
#68383 commented on
May 27, 2024 • 0 new comments -
Change grappler to optimize MatMul + AddV2 + Maximum(x,0) as FusedMatMul with Relu.
#68373 commented on
Jun 2, 2024 • 0 new comments -
tf.raw_ops.ResourceApplyGradientDescent: Aborted (core dumped)
#63695 commented on
May 30, 2024 • 0 new comments -
tf.image.draw_bounding_boxes: Aborted (core dumped)
#63688 commented on
May 30, 2024 • 0 new comments -
Segmentation fault with tf.raw_ops.AudioSpectrogram
#63352 commented on
May 30, 2024 • 0 new comments -
TFLite FP16 with Core ML Delegate gets wrong result and has no speed up
#62786 commented on
May 29, 2024 • 0 new comments -
Build Tensorflow version that detects CPU instruction set at runtime and lights-up/down
#25590 commented on
May 29, 2024 • 0 new comments -
[TFLite] Quantized int8/16 support for TILE operator
#50424 commented on
May 29, 2024 • 0 new comments -
[TFLite] Add int8 and int16x8 support for CONV_3D operator
#51527 commented on
May 29, 2024 • 0 new comments -
[TFLite] Add int8 and int16x8 support for TRANSPOSE_CONV_3D operator
#51572 commented on
May 29, 2024 • 0 new comments -
Update - Image Classification
#57046 commented on
May 29, 2024 • 0 new comments -
Update floorDiv
#58410 commented on
May 29, 2024 • 0 new comments -
Make TensorFlow compatible with PyPy
#252 commented on
May 28, 2024 • 0 new comments -
Tensorflow not detecting GPU
#64881 commented on
May 28, 2024 • 0 new comments -
Fix endianness issues in arithmetic_optimizer_test.cc tests
#59851 commented on
May 29, 2024 • 0 new comments -
Prebuilt binaries do not work with CPUs that do not have AVX instruction sets.
#19584 commented on
May 28, 2024 • 0 new comments -
Big-endian saved_model byte swapping improvements
#60356 commented on
May 29, 2024 • 0 new comments -
Search in Embedding Projector using Japanese or Hindi text causes Cannot read property 'toString' of undefined
#21891 commented on
May 28, 2024 • 0 new comments -
validate argument axis of tf.experimental.numpy.stack
#60727 commented on
May 29, 2024 • 0 new comments -
Update batch_dims argument for GatherV2
#60731 commented on
May 29, 2024 • 0 new comments -
Change TFL_MINIMUM_OS_VERSION to build TensorFlowLiteCMetal_framework on XCode 14.3
#61174 commented on
May 29, 2024 • 0 new comments -
[XLA][StreamExecutor] add empty implementation for host stream, avoid…
#61888 commented on
May 29, 2024 • 0 new comments -
update cpuinfo
#63850 commented on
May 29, 2024 • 0 new comments -
Introduce hermetic CUDA in Google ML projects.
#64130 commented on
May 30, 2024 • 0 new comments -
Inline one-off use small functions as lambdas
#64712 commented on
May 30, 2024 • 0 new comments -
Automated Code Change
#65711 commented on
May 28, 2024 • 0 new comments -
Support of C++20 for OSS TF for Linux x86_64 Ubuntu 22.04
#65948 commented on
May 30, 2024 • 0 new comments -
Update compatibility constraint in BUILD file.
#66069 commented on
Jun 1, 2024 • 0 new comments -
Introduce the MaybeBatchDown helper method, only supporting kBatchDown at this time.
#66623 commented on
May 27, 2024 • 0 new comments -
Introduce utility function GetPrevAllowedBatchSize.
#66624 commented on
May 27, 2024 • 0 new comments