Code

feat: ONNX model inference on Spark#1152

Code

Merged

mhamilton723 merged 21 commits intomicrosoft:mastermicrosoft/SynapseML:masterfrom memoryz:jasowang/onnxmemoryz/mmlspark:jasowang/onnxCopy head branch name to clipboardAug 9, 2021

+1,656-258

Commits

Commits on Aug 3, 2021

• WIP: ONNX model

  memoryz

  committed
  5dc5bbb

  Copy full SHA for 5dc5bbb

• WIP: onnx inference

  memoryz

  committed
  cc15a5d

  Copy full SHA for cc15a5d

• WIP: support input type coersion

  memoryz

  committed
  7916d07

  Copy full SHA for 7916d07

• Cleanup of resources

  memoryz

  committed
  ac59c9c

  Copy full SHA for ac59c9c

• support vector input, and convert to tensor

  memoryz

  committed
  4fc9fcc

  Copy full SHA for 4fc9fcc

• WIP: image transformer

  memoryz

  committed
  820ea69

  Copy full SHA for 820ea69

• Adding center crop

  memoryz

  committed
  1722d8c

  Copy full SHA for 1722d8c

• ONNX model unit tests

  memoryz

  committed
  5aaec34

  Copy full SHA for 5aaec34

• Adding more test cases

  memoryz

  committed
  cc327ab

  Copy full SHA for cc327ab

• Resize keep aspect ratio

  memoryz

  committed
  31da761

  Copy full SHA for 31da761

• Bug fixes

  memoryz

  committed
  7ca4071

  Copy full SHA for 7ca4071

• Unit tests

  memoryz

  committed
  e3aa223

  Copy full SHA for e3aa223

• Fixing style

  memoryz

  committed
  5cd9f3f

  Copy full SHA for 5cd9f3f

• Fixing unit test

  memoryz

  committed
  bdad5a4

  Copy full SHA for bdad5a4

• Update notebook for resize function

  memoryz

  committed
  e3518fe

  Copy full SHA for e3518fe

• Fixing code review issues

  memoryz

  committed
  0aad602

  Copy full SHA for 0aad602

Commits on Aug 4, 2021

• Add doc string to explain onnx object model

  memoryz

  committed
  fbb70a3

  Copy full SHA for fbb70a3

Commits on Aug 6, 2021

• Fixing memory leak issue

  memoryz

  committed
  0e6593e

  Copy full SHA for 0e6593e

Commits on Aug 7, 2021

• Cleanup

  memoryz

  committed
  2270830

  Copy full SHA for 2270830

Commits on Aug 9, 2021

• Support GPU

  memoryz

  committed
  e71d41e

  Copy full SHA for e71d41e

• Adding params for optimization level and device type

  memoryz

  committed
  c467152

  Copy full SHA for c467152