Revise planTensors in Linear (#5689)
author김용섭/On-Device Lab(SR)/Engineer/삼성전자 <yons.kim@samsung.com>
Fri, 19 Jul 2019 02:19:50 +0000 (11:19 +0900)
committer이한종/On-Device Lab(SR)/Engineer/삼성전자 <hanjoung.lee@samsung.com>
Fri, 19 Jul 2019 02:19:50 +0000 (11:19 +0900)
commit4d725edab14435769a2b09b51761a0ede71cfeb2
treeadea40088a2d2b5c5f3adaca7410b20181810121
parent3398f87f89dcd761b43af6df760a66e7de4f92da
Revise planTensors in Linear (#5689)

* Revise planTensors in Linear

- Introduce tensor_builder_map instead of iterTensorBuilders
- Replace the type of constants from std::vector<model::OperandIndex>
to model::OperandIndexSequence constants
- Add model::OperandIndexMap<uint32_t> def_map
- Append the step for disposing and validation

Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
* Avoid cases where insertion of unordered_map could happen
runtimes/neurun/core/src/linear/Linear.cc