ashbhandare 1c316d0e39 Parameter,Module and Optimizer changes (#11494) ... * Module step * On device training offline composition * Working grad accumulation with test for TrainStep * Temp changes * Revert "On device training offline composition" This reverts commit ec3da68247. * cleanup * Implement eval step * Use new graphs and checkpoints * Optimizer test, changes * review comments * review comments * review comments Co-authored-by: Aishwarya Bhandare <aibhanda@microsoft.com@orttrainingdev7.d32nl1ml4oruzj4qz3bqlggovf.px.internal.cloudapp.net> Co-authored-by: Baiju Meswani <bmeswani@microsoft.com>		2022-05-31 09:20:47 -07:00
..
codegen	Refactor transformers and other code to reduce memory allocation calls (#10523)	2022-02-24 16:17:14 -08:00
common	Fixed compilation error for ARM architecture (#11223)	2022-04-15 09:24:21 -07:00
dll
dlpack	Improve transfered time from ort to torch (#9610)	2022-04-06 09:12:58 +02:00
eager	Update our absl cmake files (#10762)	2022-03-04 09:28:04 -08:00
flatbuffers	Clarify usage of kOnnxDomainAlias. (#10962)	2022-03-25 09:52:59 +10:00
framework	Parameter,Module and Optimizer changes (#11494)	2022-05-31 09:20:47 -07:00
graph	Add better error message for subgraph output coming directly from outer scope value. (#11638)	2022-05-28 08:04:27 +10:00
language_interop_ops
mlas	POWER: Optimize MlasQLinearAddKernelHelper() (#11454)	2022-05-12 18:38:45 -07:00
optimizer	[CUDA] Implement BitmaskDropout, BitmaskBiasDropout and BitmaskDropoutGrad (#11534)	2022-05-27 17:24:47 +08:00
platform	Avoid using word 'crazy' (#11396)	2022-05-05 23:07:50 +08:00
providers	MTA AdamWOptimizer (#11506)	2022-05-27 19:52:04 +08:00
quantization
session	MTA AdamWOptimizer (#11506)	2022-05-27 19:52:04 +08:00
util	Upgrade emsdk to 3.1.3 (#10577)	2022-02-28 23:52:41 -08:00