mv_machine_learning: bug fix to tensor buffer allocation

[Version] : 0.26.6
[Issue type] : bug fix

Fixed a bug to wrong tensor buffer allocation since commit-id
ee852016683c6fed53fdddcc54ebea2c9212e4ff

For the output tensor, we have to allocate the output tensor buffer
not input tensor one.

Change-Id: I0653e7726666ae122e42e52d6ced8874d2b1ac3c
Signed-off-by: Inki Dae <inki.dae@samsung.com>

diff --git a/mv_machine_learning/inference/src/Inference.cpp b/mv_machine_learning/inference/src/Inference.cpp
index b6d7170..0681849 100644 (file)
--- a/mv_machine_learning/inference/src/Inference.cpp
+++ b/mv_machine_learning/inference/src/Inference.cpp
@@ -720,10 +720,10 @@ int Inference::PrepareTenosrBuffers(void)
                for (auto &layer : mOutputLayerProperty.layers) {
                        inference_engine_tensor_buffer tensor_buffer;
 
-                       ret = mInputTensorBuffers.allocate(tensor_buffer, layer.second);
+                       ret = mOutputTensorBuffers.allocate(tensor_buffer, layer.second);
                        if (ret != INFERENCE_ENGINE_ERROR_NONE) {
                                LOGE("Fail to allocate tensor buffer.");
-                               mInputTensorBuffers.release();
+                               mOutputTensorBuffers.release();
                                return ret;
                        }
 

diff --git a/packaging/capi-media-vision.spec b/packaging/capi-media-vision.spec
index 7a1db55..217969c 100644 (file)
--- a/packaging/capi-media-vision.spec
+++ b/packaging/capi-media-vision.spec
@@ -1,6 +1,6 @@
 Name:        capi-media-vision
 Summary:     Media Vision library for Tizen Native API
-Version:     0.26.5
+Version:     0.26.6
 Release:     0
 Group:       Multimedia/Framework
 License:     Apache-2.0 and BSD-3-Clause