Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Sep 15, 2023

• Remove redundant Resolve() after each inlined function (#17556)

  Show description for fdb1326
  yuslepukhin

  authored
  fdb1326

  Copy full SHA for fdb1326

Commits on Sep 8, 2023

• Update C/C++ dependencies: abseil, date, nsync, googletest, wil, mp11, cpuinfo and safeint (#15470)

  Show description for bc84f52
  

  Changming Sun

  authored
  bc84f52

  Copy full SHA for bc84f52

• Updates to training pipelines (#17292)

  askhade

  authored
  c5dbd5c

  Copy full SHA for c5dbd5c

Commits on Sep 6, 2023

• [ORTModule] Add Manual Seed to Fix UT Failure (#17411)

  Show description for deda5db
  Lafi7e

  authored
  deda5db

  Copy full SHA for deda5db

Commits on Sep 1, 2023

• Change RuntimeError to ImportError (#17380)

  Show description for 8b98eca
  baijumeswani

  authored
  8b98eca

  Copy full SHA for 8b98eca

Commits on Aug 31, 2023

• Fuse ScaledSum and its backward BatchScale (#16517)

  Show description for 58af36b
  pengwa

  authored
  58af36b

  Copy full SHA for 58af36b

Commits on Aug 29, 2023

• add logits option to generate artifacts (#17276)

  Show description for 8224891
  AdamLouly

  authored
  8224891

  Copy full SHA for 8224891

• EvalStep called with wrong inputs onnxruntime_training_cxx_inline.h (#17331)

  kushalpatil07

  authored
  7b92057

  Copy full SHA for 7b92057

• Increase max error tolerance for ConvTransposeGrad test (#17315)

  baijumeswani

  authored
  38ea8c3

  Copy full SHA for 38ea8c3

Commits on Aug 25, 2023

• Add support for more ops for padding elimination (#17217)

  Show description for 401129d
  guyang3532

  authored
  401129d

  Copy full SHA for 401129d

Commits on Aug 24, 2023

• Introduce ZeROOffloadSubscriber for ORTModule (#17006)

  Show description for d90afc6
  pengwa

  authored
  d90afc6

  Copy full SHA for d90afc6

• ConvTransposeGrad CUDA Kernel (#17201)

  baijumeswani

  authored
  fca81cc

  Copy full SHA for fca81cc

• Build gradient graph starting at the loss alone (#17240)

  baijumeswani

  authored
  34d18ee

  Copy full SHA for 34d18ee

Commits on Aug 23, 2023

• On-Device Training - Enable loading from buffer (#16417)

  askhade

  authored
  56102ec

  Copy full SHA for 56102ec

Commits on Aug 18, 2023

• Add SGDOptimizer in the on-device training offline tooling (onnxblock) (#17085)

  Show description for c0b6c6c
  AdamLouly

  authored
  c0b6c6c

  Copy full SHA for c0b6c6c

• Move some tests from CUDA only to CPU (#17189)

  Show description for 68a670c
  askhade

  authored
  68a670c

  Copy full SHA for 68a670c

Commits on Aug 16, 2023

• NVTX windows include and link fixes (#16831)

  Show description for 7b9d1f1
  gedoensmax

  authored
  7b9d1f1

  Copy full SHA for 7b9d1f1

Commits on Aug 15, 2023

• PythonOp Enhancement: Bool and Tuple[Bool] Constants, Materialize Grads, Empty Inputs, Save In Context (#16828)

  Show description for abf9765
  pengwa

  authored
  abf9765

  Copy full SHA for abf9765

Commits on Aug 14, 2023

• Allow defining customized PythonOp shape inferer (#17093)

  Show description for cd7b3f5
  pengwa

  authored
  cd7b3f5

  Copy full SHA for cd7b3f5

Commits on Aug 11, 2023

• LeakyRelu Gradient (#17039)

  baijumeswani

  authored
  3e7f70b

  Copy full SHA for 3e7f70b

Commits on Aug 10, 2023

• Update model_tests.cc: avoid auto adding new tests from new opsets (#17084)

  Show description for 6dffd1a
  

  Changming Sun

  authored
  6dffd1a

  Copy full SHA for 6dffd1a

• [ROCm] Update CI based on ubuntu 22.04 (#17076)

  Show description for 12837ba
  PeixuanZuo

  authored
  12837ba

  Copy full SHA for 12837ba

• support broadcast shape for elementwise node in padding elimination (#16710)

  Show description for ef6f4a4
  guyang3532

  authored
  ef6f4a4

  Copy full SHA for ef6f4a4

• Check type for building gradient graph (#17046)

  Show description for 0471f6f
  pengwa

  authored
  0471f6f

  Copy full SHA for 0471f6f

• GRU Training and GRU Gradient Kernels (#16929)

  baijumeswani

  authored
  31cbd63

  Copy full SHA for 31cbd63

• Copy to buffer for both trainable as well as non trainable parameters (#17070)

  baijumeswani

  authored
  f17efb5

  Copy full SHA for f17efb5

Commits on Aug 9, 2023

• [CUDA][ROCm] Allow allocating ScratchBuffer from TuningContext (#17028)

  Show description for a4902ee
  cloudhan

  authored
  a4902ee

  Copy full SHA for a4902ee

• Use full qualified name for PythonOp export (#17021)

  Show description for 6e6f582
  pengwa

  authored
  6e6f582

  Copy full SHA for 6e6f582

Commits on Aug 8, 2023

• Fix orttraining_test_dort.py (#17034)

  Show description for 45ea907
  titaiwangms

  authored
  45ea907

  Copy full SHA for 45ea907

• Add mac and windows python packages for onnxruntime-training (#16993)

  baijumeswani

  authored
  249917a

  Copy full SHA for 249917a

Commits on Aug 7, 2023

• Update torch.onnx.OnnxRegistry usage in DORT tests (#17009)

  Show description for 8a335b8
  

  titaiwangms

  and

  wschin

  authored
  8a335b8

  Copy full SHA for 8a335b8

• Fix few small bugs (#17019)

  Show description for 3649376
  pengwa

  authored
  3649376

  Copy full SHA for 3649376

Commits on Aug 4, 2023

• Add Gradient for Reciprocal (#16945)

  baijumeswani

  authored
  e5bb7ab

  Copy full SHA for e5bb7ab

• Refactor schema extraction and output unflattening (#16894)

  Show description for a6887f1
  pengwa

  authored
  a6887f1

  Copy full SHA for a6887f1

• [CoreML EP] Enable inputs with dynamic shape (#16915)

  Show description for f98d3f8
  edgchen1

  authored
  f98d3f8

  Copy full SHA for f98d3f8