Commits

Branch selector

 bc4b9d8

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 29, 2024

• Enhanced zeta readability based on the article

  Show description for bc4b9d8
  

  Paweł Paruzel

  authored and

  tensorflower-gardener

  committed
  bc4b9d8

  Copy full SHA for bc4b9d8

• [odml] Remove MHLO from CHLO->StableHLO lowering. Migrate JAX random lowering to StableHLO

  Show description for b36854a
  

  GleasonK

  authored and

  tensorflower-gardener

  committed
  b36854a

  Copy full SHA for b36854a

• Restore GOOGLE_CUDA guard in scoped_annotation.h

  Show description for 210d156
  tensorflower-gardener

  committed
  210d156

  Copy full SHA for 210d156

• Delete the redundant compilation_cache_test

  Show description for 370c98d
  tensorflower-gardener

  committed
  370c98d

  Copy full SHA for 370c98d

• Deduplicate inferred mesh shapes when try_multiple_mesh_shapes=true.

  Show description for bcf3638
  tensorflower-gardener

  committed
  bcf3638

  Copy full SHA for bcf3638

• Split hybrid quantized dot-like StableHLO ops into TFLite dequantize and float op

  Show description for 147638c
  

  doyeonkim0

  authored and

  tensorflower-gardener

  committed
  147638c

  Copy full SHA for 147638c

• [xla:gpu][NFC] Use meaningful constexpr

  Show description for 06f2799
  

  tyb0807

  authored and

  tensorflower-gardener

  committed
  06f2799

  Copy full SHA for 06f2799

• Update GraphDef version to 1816.

  Show description for 3016662
  tensorflower-gardener

  committed
  3016662

  Copy full SHA for 3016662

• compat: Update forward compatibility horizon to 2024-03-29

  Show description for 17aaaa5
  tensorflower-gardener

  committed
  17aaaa5

  Copy full SHA for 17aaaa5

• [xla:gpu][NFC] Add AddressComputationThunk test with GEMM operands sharing the same buffer

  Show description for f2a78c0
  

  tyb0807

  authored and

  tensorflower-gardener

  committed
  f2a78c0

  Copy full SHA for f2a78c0

• Convert quantized stablehlo.constant to tfl.pseudo_qconst

  Show description for 06ef8a6
  

  doyeonkim0

  authored and

  tensorflower-gardener

  committed
  06ef8a6

  Copy full SHA for 06ef8a6

• [xla:gpu] Create fake buffer allocations for embedded thunk

  Show description for a829ac0
  

  tyb0807

  authored and

  tensorflower-gardener

  committed
  a829ac0

  Copy full SHA for a829ac0

• [xla][gpu] Extracting triton codegen requirements for hlo instructions

  Show description for 3d9b57d
  tensorflower-gardener

  committed
  3d9b57d

  Copy full SHA for 3d9b57d

• [xla:gpu][NFC] Use absl::Span more consistenly

  Show description for 903f741
  

  tyb0807

  authored and

  tensorflower-gardener

  committed
  903f741

  Copy full SHA for 903f741

• Automated Code Change

  Show description for 9bad4ef
  tensorflower-gardener

  committed
  9bad4ef

  Copy full SHA for 9bad4ef

• Go: Update generated wrapper functions for TensorFlow ops.

  Show description for 8330127
  tensorflower-gardener

  committed
  8330127

  Copy full SHA for 8330127

• Integrate LLVM at llvm/llvm-project@aa2c14de1adc

  Show description for 0a9c747
  

  gribozavr

  authored and

  tensorflower-gardener

  committed
  0a9c747

  Copy full SHA for 0a9c747

• [XLA] Respect min_rank for reduce scatter version of MatchReduceScatter.

  Show description for e84bf19
  

  Marcello Maggioni

  authored and

  tensorflower-gardener

  committed
  e84bf19

  Copy full SHA for e84bf19

• Enable weight-only quantization of stablehlo.convolution

  Show description for 2ec2502
  

  doyeonkim0

  authored and

  tensorflower-gardener

  committed
  2ec2502

  Copy full SHA for 2ec2502

• Add missing 'END' to definition of XlaSplitND.

  Show description for a6bbcb7
  

  matthiaskramm

  authored and

  tensorflower-gardener

  committed
  a6bbcb7

  Copy full SHA for a6bbcb7

• Set release_base for all release platforms

  Show description for 41bfc5d
  

  jakeharmon8

  authored and

  tensorflower-gardener

  committed
  41bfc5d

  Copy full SHA for 41bfc5d

• Remove the duplicate device assignment in ifrt_serving_executable.

  Show description for 56e98e5
  

  SiqiaoWu1993

  authored and

  tensorflower-gardener

  committed
  56e98e5

  Copy full SHA for 56e98e5

• Correctly handle output streaming case where the MoveToHost annotation is the entry computation root

  Show description for 7fadd47
  tensorflower-gardener

  committed
  7fadd47

  Copy full SHA for 7fadd47

• Add a Resource for KV Cache buffer storage

  Show description for 90ab9c1
  

  talumbau

  authored and

  tensorflower-gardener

  committed
  90ab9c1

  Copy full SHA for 90ab9c1

• Delete populateRankSpecialization*Patterns functions

  Show description for 8ae67af
  

  Michael Levesque-Dion

  authored and

  tensorflower-gardener

  committed
  8ae67af

  Copy full SHA for 8ae67af

• Internal cleanup of BUILD/.bzl files

  Show description for a3db0ab
  tensorflower-gardener

  committed
  a3db0ab

  Copy full SHA for a3db0ab

Commits on Mar 28, 2024

• Move sparsecore passes under transforms/sparsecore.

  Show description for 0adcfc2
  

  matthiaskramm

  authored and

  tensorflower-gardener

  committed
  0adcfc2

  Copy full SHA for 0adcfc2

• Integrate StableHLO at openxla/stablehlo@271e8634

  Show description for a9db13b
  

  ghpvnist

  authored and

  tensorflower-gardener

  committed
  a9db13b

  Copy full SHA for a9db13b

• [XLA:Python] Adding `xla::PrimitiveType` <-> `numpy.dtype` conversions to the library for internal debugging tools.

  Show description for a7d81b5
  

  wrengr

  authored and

  tensorflower-gardener

  committed
  a7d81b5

  Copy full SHA for a7d81b5

• Add quantization/legalization for `stablehlo.add` and respective pipeline changes.

  Show description for 9926a40
  

  jenriver

  authored and

  tensorflower-gardener

  committed
  9926a40

  Copy full SHA for 9926a40

• Reduce memory usage of convert_control_to_data_outputs.

  Show description for 2d48a6d
  

  matthiaskramm

  authored and

  tensorflower-gardener

  committed
  2d48a6d

  Copy full SHA for 2d48a6d

• [XLA:Python] Add python function to convert `xla::LiteralProto` into a tuple-tree of `numpy.ndarray`.

  Show description for 86e08a6
  

  wrengr

  authored and

  tensorflower-gardener

  committed
  86e08a6

  Copy full SHA for 86e08a6

• [MLIR Exporter] Move created `NodeDef` and `AttrValue` protocol buffers into the exported `Graph`.

  Show description for 3107f40
  

  mrry

  authored and

  tensorflower-gardener

  committed
  3107f40

  Copy full SHA for 3107f40

• Add a field to StreamZ metric /tensorflow/core/tf_mlir_bridge_first_phase_count

  Show description for d952cc2
  

  jcai19

  authored and

  tensorflower-gardener

  committed
  d952cc2

  Copy full SHA for d952cc2

• Add support for tiled sharding with replicate_on_last_tile_dim in TpuRewrite pass

  Show description for 97fd9c1
  

  swachhandl

  authored and

  tensorflower-gardener

  committed
  97fd9c1

  Copy full SHA for 97fd9c1

Pagination

Previous

Next