Commits

Branch selector

 master

User selector

dayeongl

Datepicker

All time

Commit History

Commits on Jan 17, 2022

• Add optimized kernel for int16 sub op for AVX2.

  Show description for c93e475
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  c93e475

  Copy full SHA for c93e475

Commits on Jan 13, 2022

• Add optimized kernel for int16 leaky_relu op for AVX2.

  Show description for 6d33f19
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  6d33f19

  Copy full SHA for 6d33f19

Commits on Jan 12, 2022

• Add optimized kernel for int16 add op for AVX2.

  Show description for 939011a
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  939011a

  Copy full SHA for 939011a

Commits on Jan 8, 2022

• Add optimized kerenel for int16 add op for Neon.

  Show description for ffc99d0
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  ffc99d0

  Copy full SHA for ffc99d0

Commits on Jan 7, 2022

• Add optimized kerenel for int16 add op for Neon.

  Show description for a7fddbd
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  a7fddbd

  Copy full SHA for a7fddbd

Commits on Jan 5, 2022

• Support Conv2DTranspose for 16x8 quant with 32bit bias.

  Show description for 0862854
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  0862854

  Copy full SHA for 0862854

Commits on Jan 3, 2022

• Support Conv2D for 16x8 quant with 32bit bias.

  Show description for 0d8705c
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  0d8705c

  Copy full SHA for 0d8705c

Commits on Dec 30, 2021

• Add option to store bias as 32 bit in 16x8 Quant.

  Show description for ea33c1e
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  ea33c1e

  Copy full SHA for ea33c1e

Commits on Nov 8, 2021

• Update TFLite kernel to use Ruy 16x8 Gemm instead of reference kernel.

  Show description for 5de1820
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  5de1820

  Copy full SHA for 5de1820

Commits on Nov 4, 2021

• Update cache_policy for fully_connected:integer_ops.

  Show description for 07a7a5e
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  07a7a5e

  Copy full SHA for 07a7a5e

Commits on Aug 11, 2021

• Use max_execution_batch_size() instead of options.max_execution_batch_size.

  Show description for 9b01d1f
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  9b01d1f

  Copy full SHA for 9b01d1f

Commits on Mar 6, 2020

• Support user-specified string input in tflite benchmark tools.

  Show description for 70fab67
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  70fab67

  Copy full SHA for 70fab67

Commits on Mar 4, 2020

• Add tests to check the initialization of the input tensor.

  Show description for 5731c94
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  5731c94

  Copy full SHA for 5731c94

Commits on Feb 27, 2020

• Custom input support for simply file-based input.

  Show description for 70cdf91
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  70cdf91

  Copy full SHA for 70cdf91

• Custom input support for simply file-based input.

  Show description for f5a6404
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  f5a6404

  Copy full SHA for f5a6404

Commits on Feb 21, 2020

• Fix ProfilingListener for subclasses to override.

  Show description for e38ef04
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  e38ef04

  Copy full SHA for e38ef04

Commits on Feb 19, 2020

• Move the writer functions of profileSummarizer to ProfileSummaryFormatter.

  Show description for ee7642b
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  ee7642b

  Copy full SHA for ee7642b

Commits on Feb 18, 2020

• Moves ProfilingListener to "profiling_listener.h", "profiling_listener.cc" for InternalBenchmarkTfLiteModel to instantiate this class.

  Show description for 1efab01
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  1efab01

  Copy full SHA for 1efab01

Commits on Feb 13, 2020

• Update stats_calculator to fix csv format error in node name

  Show description for 8cad6f5
  

  dayeongl

  authored and

  tensorflower-gardener

  committed
  8cad6f5

  Copy full SHA for 8cad6f5