projects / platform / core / ml / nnfw.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 0c0530e) | patch
[neurun] Do not collect TensorBuilders one by one (#6458)
author이한종/On-Device Lab(SR)/Engineer/삼성전자 <hanjoung.lee@samsung.com>
Mon, 12 Aug 2019 10:34:39 +0000 (19:34 +0900)
committer오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 <hseok82.oh@samsung.com>
Mon, 12 Aug 2019 10:34:39 +0000 (19:34 +0900)
commit43a5d8aba6b8ed3cbd7cb441741294672d345db9
tree8de4bc4ca9f037ede683b55f5990d58d90f966dctree | snapshot
parent0c0530e81133d660c1accbb090d8a527f622eae7commit | diff
[neurun] Do not collect TensorBuilders one by one (#6458)

* [neurun] Do not collect TensorBuilders one by one

To get the full set of tensor builders currently using it used to
collect from all the `LowerInfo`s. Now that `BackendResolver` manages
them all we can just use it.
Allow allocation of 0-byte object

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
runtimes/neurun/backend/cpu/MemoryPlanner.cc diff | blob | history
runtimes/neurun/core/src/compiler/BackendResolver.h diff | blob | history
runtimes/neurun/core/src/compiler/ExecutorFactory.cc diff | blob | history
runtimes/neurun/core/src/compiler/Linear.cc diff | blob | history
runtimes/neurun/core/src/compiler/Linear.h diff | blob | history
Domain: Machine Learning / ML Framework;