Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Dec 19, 2019

• Update NNPI Backend to v0.4.1.7 (pytorch#3920)

  Show description for 1e7a5f9
  

  omromano

  authored and

  facebook-github-bot

  committed
  1e7a5f9

  Copy full SHA for 1e7a5f9

• Add ReluGrad instruction

  Show description for 17b326f
  

  Meghan Lele

  authored and

  facebook-github-bot

  committed
  17b326f

  Copy full SHA for 17b326f

• Add Model Tuner front-end tool. (pytorch#3816)

  Show description for 41d3e30
  

  mciprian13

  authored and

  facebook-github-bot

  committed
  41d3e30

  Copy full SHA for 41d3e30

• Implement LocalResponseNormalization{,Grad}Inst (pytorch#3814)

  Show description for de133ee
  

  pjaaskel

  authored and

  facebook-github-bot

  committed
  de133ee

  Copy full SHA for de133ee

Commits on Dec 18, 2019

• add method for updating a device Tensor without copying it to host (pytorch#3923)

  Show description for ba494df
  

  nickgg

  authored and

  facebook-github-bot

  committed
  ba494df

  Copy full SHA for ba494df

• Use Node names for vertices in dotty DAG dump (pytorch#3924)

  Show description for 9f63262
  

  nickgg

  authored and

  facebook-github-bot

  committed
  9f63262

  Copy full SHA for 9f63262

• Improve some lowered and auto differentiated Node names (pytorch#3918)

  Show description for 3c685f6
  

  nickgg

  authored and

  facebook-github-bot

  committed
  3c685f6

  Copy full SHA for 3c685f6

• Let repro dump its outputs (pytorch#3894)

  Show description for e9b42bb
  

  Yinghai Lu

  authored and

  facebook-github-bot

  committed
  e9b42bb

  Copy full SHA for e9b42bb

• fix ReduceSum/ReduceMean outout in ONNXModelWriter

  Show description for a1c1ce2
  

  tracelogfb

  authored and

  facebook-github-bot

  committed
  a1c1ce2

  Copy full SHA for a1c1ce2

• Add clipping only inputs option and clip + convert Storage (pytorch#3855)

  Show description for aab4944
  

  jfix71

  authored and

  facebook-github-bot

  committed
  aab4944

  Copy full SHA for aab4944

• Fix TensorView offsets to account for strides (pytorch#3922)

  Show description for cb46537
  

  jfix71

  authored and

  facebook-github-bot

  committed
  cb46537

  Copy full SHA for cb46537

Commits on Dec 17, 2019

• Modify OptimizeArithmeticNodes to correctly handle multi-result nodes

  Show description for 4683f3d
  

  Meghan Lele

  authored and

  facebook-github-bot

  committed
  4683f3d

  Copy full SHA for 4683f3d

Commits on Dec 16, 2019

• Adding unfused SLS

  Show description for fd70461
  

  mjanderson09

  authored and

  facebook-github-bot

  committed
  fd70461

  Copy full SHA for fd70461

• Add ability to load sentences for translation, and check expected result sentences from input files (pytorch#3907)

  Show description for bedd9c0
  

  jfix71

  authored and

  facebook-github-bot

  committed
  bedd9c0

  Copy full SHA for bedd9c0

• Use dim_t instead of size_t (pytorch#3913)

  Show description for bdb74bd
  

  jackm321

  authored and

  facebook-github-bot

  committed
  bdb74bd

  Copy full SHA for bdb74bd

Commits on Dec 15, 2019

• Add ctr_mbl_feed test and fix bugs (pytorch#3912)

  Show description for 19988e7
  

  jackm321

  authored and

  facebook-github-bot

  committed
  19988e7

  Copy full SHA for 19988e7

Commits on Dec 14, 2019

• Disabel GroupwiseQuantizedConvolution/0 for OpenCL backend (pytorch#3909)

  Show description for 301e10f
  

  Yinghai Lu

  authored and

  facebook-github-bot

  committed
  301e10f

  Copy full SHA for 301e10f

• Add flag to disable NNPI Private transforms (pytorch#3910)

  Show description for acc4de8
  

  nrsatish

  authored and

  facebook-github-bot

  committed
  acc4de8

  Copy full SHA for acc4de8

• Enable channelwise quantization of conv2d & fix some bugs (pytorch#3877)

  Show description for 66eb320
  

  zrphercule

  authored and

  facebook-github-bot

  committed
  66eb320

  Copy full SHA for 66eb320

• Compile PyTorch model ahead of time (pytorch#3902)

  Show description for 5f0ee4d
  

  Qi Zhou

  authored and

  facebook-github-bot

  committed
  5f0ee4d

  Copy full SHA for 5f0ee4d

• Make NNPI memory settable for ICEREF (pytorch#3906)

  Show description for b8f90a7
  

  jfix71

  authored and

  facebook-github-bot

  committed
  b8f90a7

  Copy full SHA for b8f90a7

Commits on Dec 13, 2019

• Remove unused NNPIBackendTestUtils.h (pytorch#3905)

  Show description for 60f65f2
  

  jfix71

  authored and

  facebook-github-bot

  committed
  60f65f2

  Copy full SHA for 60f65f2

• Lower all BatchMatMul by default, add cl opt to disable (pytorch#3904)

  Show description for 76218ee
  

  jfix71

  authored and

  facebook-github-bot

  committed
  76218ee

  Copy full SHA for 76218ee

• Legalize bundle name. Change Loader default backend to CPU. (pytorch#3901)

  Show description for 0dd6ff2
  

  mciprian13

  authored and

  facebook-github-bot

  committed
  0dd6ff2

  Copy full SHA for 0dd6ff2

• Add SLWS in the interpreter for fp16 (pytorch#3896)

  Show description for 5faf60f
  

  SameerAsal

  authored and

  facebook-github-bot

  committed
  5faf60f

  Copy full SHA for 5faf60f

• Add RNN and GRU. (pytorch#3847)

  Show description for 306a648
  

  mciprian13

  authored and

  facebook-github-bot

  committed
  306a648

  Copy full SHA for 306a648

• Unsilo Interpreter Backend, use directly by GraphOptimizer to allow other backends to call passes directly, and wrap transformPostLowering in if again (pytorch#3900)

  Show description for 05144c0
  

  jfix71

  authored and

  facebook-github-bot

  committed
  05144c0

  Copy full SHA for 05144c0

Commits on Dec 12, 2019

• Various fixes (pytorch#3895)

  Show description for 3e8854f
  

  jackm321

  authored and

  facebook-github-bot

  committed
  3e8854f

  Copy full SHA for 3e8854f

• Fix build (size_t -> dim_t) (pytorch#3899)

  Show description for 4bfe34e
  

  jfix71

  authored and

  facebook-github-bot

  committed
  4bfe34e

  Copy full SHA for 4bfe34e

• Add removeClipsBlockingFusion pass (pytorch#3886)

  Show description for 8a6b141
  

  jfix71

  authored and

  facebook-github-bot

  committed
  8a6b141

  Copy full SHA for 8a6b141

• Add delayed NNPI lowering of FP32 BMM (pytorch#3890)

  Show description for a0f8e3b
  

  jfix71

  authored and

  facebook-github-bot

  committed
  a0f8e3b

  Copy full SHA for a0f8e3b

• Update NNPI Backend to v0.4.1.2 (pytorch#3879)

  Show description for 9633d96
  

  omromano

  authored and

  facebook-github-bot

  committed
  9633d96

  Copy full SHA for 9633d96

• fix potential use after free / segfault in provisioner (pytorch#3897)

  Show description for 62d27c6
  

  tracelogfb

  authored and

  facebook-github-bot

  committed
  62d27c6

  Copy full SHA for 62d27c6

• Allow user to resolve undefined model symbols ("dim_param"). (pytorch#3817)

  Show description for af37983
  

  mciprian13

  authored and

  facebook-github-bot

  committed
  af37983

  Copy full SHA for af37983

Commits on Dec 11, 2019

• Add CompilationContext for manual partitioner (pytorch#3887)

  Show description for e9cd3c9
  

  nrsatish

  authored and

  facebook-github-bot

  committed
  e9cd3c9

  Copy full SHA for e9cd3c9

Pagination

Previous

Next