[MachineLearning.Train] Change internal logic to resolve a memory leak
authorSeoHyungjun <hyungjun.seo@samsung.com>
Fri, 4 Oct 2024 08:17:00 +0000 (17:17 +0900)
committerSW Kim <sw0312.kim@samsung.com>
Tue, 15 Oct 2024 07:50:58 +0000 (16:50 +0900)
commit471c05d5e87a84ae63125c79647a00dcc2b3a5c2
treef67b2f16001595bc60ee82980c7e91a7b5ecbf53
parentda187aec6d088f72edeff248630bb3828258d47e
[MachineLearning.Train] Change internal logic to resolve a memory leak

There were issues with heap use after free and heap overflow occurring
(TNINE-4152, 4153, 4154).

In order to solve this problem, modified the internal logic of the API.

Signed-off-by: SeoHyungjun <hyungjun.seo@samsung.com>
src/Tizen.MachineLearning.Train/Tizen.MachineLearning.Train/Dataset.cs
src/Tizen.MachineLearning.Train/Tizen.MachineLearning.Train/Layer.cs
src/Tizen.MachineLearning.Train/Tizen.MachineLearning.Train/Model.cs
src/Tizen.MachineLearning.Train/Tizen.MachineLearning.Train/Optimizer.cs