Commits

Branch selector

 master

User selector

jianyuh

Datepicker

All time

Commit History

Commits on Aug 7, 2019

• Remove K and N function arguments for fbgemm_pack_quantized_matrix (pytorch#22956)

  Show description for 2635b62
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  2635b62

  Copy full SHA for 2635b62

• Change fbgemm_linear_{int8,fp16}_weight to fbgemm_linear_{int8,fp16}_weight_fp32_activation (pytorch#22955)

  Show description for 78cc9b9
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  78cc9b9

  Copy full SHA for 78cc9b9

Commits on Jul 26, 2019

• Add fbgemm_qlinear_dynamic op (pytorch#23104)

  Show description for 47a5429
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  47a5429

  Copy full SHA for 47a5429

Commits on Jul 25, 2019

• Fix a tiny bug and refactor (pytorch#22808)

  Show description for 39de49c
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  39de49c

  Copy full SHA for 39de49c

Commits on Jul 17, 2019

• qlinear operator level benchmark (pytorch#22914)

  Show description for f72d754
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  f72d754

  Copy full SHA for f72d754

Commits on Jul 12, 2019

• Add the support of handle Bias being nullptr for torch.ops.quantized.fbgemm_linear (pytorch#22403)

  Show description for 8ec712d
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  8ec712d

  Copy full SHA for 8ec712d

Commits on Jul 5, 2019

• Add the support of handle Bias being nullptr for torch.ops.quantized.fbgemm_conv (pytorch#22472)

  Show description for 4ba1c4f
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  4ba1c4f

  Copy full SHA for 4ba1c4f

Commits on Jun 4, 2019

• Add quantized::fbgemm_linear_unpack operator for serialization (pytorch#97)

  Show description for 0f58d20
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  0f58d20

  Copy full SHA for 0f58d20

Commits on Jun 3, 2019

• Rename FC to Linear for the function name (pytorch#21268)

  Show description for 37fed9b
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  37fed9b

  Copy full SHA for 37fed9b

• Replace AT_ASSERTM with TORCH_CHECK (pytorch#21267)

  Show description for 63b3c5a
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  63b3c5a

  Copy full SHA for 63b3c5a

Commits on May 22, 2019

• Rename FC to Linear in the test routine (pytorch#20716)

  Show description for 4a85e79
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  4a85e79

  Copy full SHA for 4a85e79

Commits on May 21, 2019

• Change Bias to QTensor with qint32(int32_t) (pytorch#20713)

  Show description for e6f22e1
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  e6f22e1

  Copy full SHA for e6f22e1

• Change Weight to QTensor with qint8(int8_t) (pytorch#20712)

  Show description for b9a150e
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  b9a150e

  Copy full SHA for b9a150e

• Fix a bug in quantize_linear (pytorch#20711)

  Show description for ac2314f
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  ac2314f

  Copy full SHA for ac2314f

Commits on May 8, 2019

• Quantized FCRelu operator (pytorch#19750)

  Show description for 4a086f7
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  4a086f7

  Copy full SHA for 4a086f7

• Use QTensor with quantized FC operator (pytorch#19541)

  Show description for e8cdfb5
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  e8cdfb5

  Copy full SHA for e8cdfb5

• Quantized FC operator (pytorch#19497)

  Show description for 1deb8dd
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  1deb8dd

  Copy full SHA for 1deb8dd

Commits on Apr 29, 2019

• Avoid printing debug information in the test

  Show description for ff33c8c
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  ff33c8c

  Copy full SHA for ff33c8c

Commits on Mar 22, 2019

• Fix alignment issues for Fake BFP16 fp32 -> bfp16 rounding routines (pytorch#18321)

  Show description for 18a6781
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  18a6781

  Copy full SHA for 18a6781

Commits on Mar 16, 2019

• Typo fix (pytorch#18089)

  Show description for 96fe2b4
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  96fe2b4

  Copy full SHA for 96fe2b4

Commits on Jan 12, 2019

• Make it consistent for OperatorBase usage (pytorch#15908)

  Show description for 264e16b
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  264e16b

  Copy full SHA for 264e16b

Commits on Jan 3, 2019

• Unify the usage of Dequantize (pytorch#15685)

  Show description for 3b5a940
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  3b5a940

  Copy full SHA for 3b5a940

Commits on Dec 19, 2018

• race condition fix of using mutable_data inside OPENMP region for batched matmul (pytorch#15371)

  Show description for cd8dd49
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  cd8dd49

  Copy full SHA for cd8dd49

Commits on Nov 9, 2018

• Change all namespace fbgemm2 in the new fbgemm2 to namespace fbgemm (pytorch#13740)

  Show description for 2ee4ef5
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  2ee4ef5

  Copy full SHA for 2ee4ef5

• Change all namespace fbgemm in the old fbgemm to namespace fbgemm0 (pytorch#13701)

  Show description for 55964ab
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  55964ab

  Copy full SHA for 55964ab