Commits

Branch selector

 master

User selector

liufengdb

Datepicker

All time

Commit History

Commits on Oct 26, 2021

• Legalize mhlo::reduce with logical_or to tf.Any

  Show description for 035ce30
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  035ce30

Commits on Oct 5, 2021

• Remove the tfl.softmax rank range [1, 4] constraint

  Show description for d91a3e7
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  d91a3e7

Commits on Jul 15, 2021

• Add auto-generated TensorFlow XlaRemoveDynamicDimensionSize op

  Show description for ab330d3
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  ab330d3

Commits on Jun 18, 2021

• Improve the generated tfr methods

  Show description for 0d9d692
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  0d9d692

Commits on Jun 2, 2021

• Move the fake quant related tests to a seperate file

  Show description for 68961af
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  68961af

Commits on May 27, 2021

• [XLA:GPU] Allow int8,int8->int32 GEMMs on GPU via cuBLAS

  Show description for a1e26cf
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  a1e26cf

Commits on May 24, 2021

• Remove a redundant dep in the llvm_gpu_backend

  Show description for 23fbdd1
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  23fbdd1

Commits on May 21, 2021

• Disable the test on mac

  Show description for 01c98e0
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  01c98e0

• Skip the default quantization parameters inside the custom op

  Show description for ed56215
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  ed56215

• Add a graph rewrite pass for adaptive batch scheduler and customize batch-op from reserved attributes.

  Show description for fcacce5
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  fcacce5

Commits on May 20, 2021

• Add CompositeTensor as a base class to tf.linalg.LinearOperator.

  Show description for e64ad43
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  e64ad43

• PR tensorflow#49108: [ROCm] Adding complex types to existing hipsparse functionality

  Show description for 875e383
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  875e383

Commits on May 11, 2021

• Add a new quantize rewrite pattern in case the input isn't quantizable

  Show description for 2146cba
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  2146cba

Commits on May 10, 2021

• Improve the tfr translator to handle more integer types and tfr quant built-in ops

  Show description for b99fce7
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  b99fce7

Commits on May 7, 2021

• Replaces all the use of the argument before it is replaced

  Show description for 57cdb45
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  57cdb45

Commits on May 6, 2021

• Add quantization related TFR built-in ops

  Show description for 1bebe43
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  1bebe43

Commits on Apr 29, 2021

• [tf-quant] Add TF compositions for the floating-point and integer tfl.conv2d

  Show description for 8d78df9
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  8d78df9

Commits on Apr 24, 2021

• Add a pass to push down the dequantize to be after the pass through op with

  Show description for 930b14b
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  930b14b

Commits on Apr 23, 2021

• Move some lambda functions to the top level so they can be reused

  Show description for 254b149
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  254b149

• Remove one outdated TODO

  Show description for a0b7d66
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  a0b7d66

• Removed one outdated TODO

  Show description for ddd1129
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  ddd1129

Commits on Apr 22, 2021

• Remove an unused pass in the tflite converter

  Show description for 88206f6
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  88206f6

• Add the tfl.if op

  Show description for e0d2ba7
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  e0d2ba7

• Remove the BuildWhileOperator for TF::WhileOp in the flatbuffer exporter

  Show description for 31f5833
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  31f5833

Commits on Apr 20, 2021

• Remove the outdated TODOs

  Show description for 91b87a8
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  91b87a8

• Merge the tf.FakeQuant* unwrapping logic and inserting QDQ logic into the same step

  Show description for 9505d23
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  9505d23

Commits on Apr 13, 2021

• Support the input/output data type inside the mlir quantizer

  Show description for fc94423
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  fc94423

Commits on Apr 5, 2021

• Merge the tf.FakeQuant* unwrapping logic and inserting QDQ logic into the same step

  Show description for a8fbda3
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  a8fbda3

• Merge the LowerCustomOps pass and the PrepareTf pass

  Show description for add13d4
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  add13d4

Commits on Apr 1, 2021

• Fix the RaiseCustomOps pass to handle ops wrapped by the tf_executor dialect

  Show description for 32355e5
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  32355e5

Commits on Mar 30, 2021

• Prevent the tf.FakeQuant* ops being folded

  Show description for 0e22131
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  0e22131

Commits on Mar 29, 2021

• Wrap tf.FakeQuantWithMinMaxVars before the TF standard legalization pipeline

  Show description for 7317647
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  7317647

Commits on Mar 24, 2021

• Convert all the constant to standard constant before quantization passes

  Show description for da3da1e
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  da3da1e

Commits on Mar 23, 2021

• Move the constant quantization patterns to a different batch

  Show description for 1ad6b35
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  1ad6b35

Commits on Mar 15, 2021

• Fix the quantization passes to handle the custom op correctly

  Show description for bbb554f
  

  liufengdb

  authored and

  tensorflower-gardener

  committed
  bbb554f

Pagination

PreviousNext