Update description of inference_engine_tensor_info structure

Change-Id: Ibd70368f453a9dea5eb5ab85c2730789428a555f
Signed-off-by: Inki Dae <inki.dae@samsung.com>

diff --git a/include/inference_engine_type.h b/include/inference_engine_type.h
index ab6a472aac5a35fadad264b73cac7f986fa846c7..d14139927bcb56cf47f906f1df0c031d5736de5e 100644 (file)
--- a/include/inference_engine_type.h
+++ b/include/inference_engine_type.h
@@ -157,6 +157,10 @@ typedef struct _inference_engine_tensor_buffer {
  *    - a name of a given layer
  *    - a tensor shape of the layer
  *    - a tensor type of the layer
+ *    - a tensor size of the layer.
+ *
+ * Caution. Tensor size is not in-memory buffer size in bytes so based on a given tensor size,
+ * upper framework should allocate actual tensor buffer according to tensor data types (i.e., uint8, float32...)
  *
  * @since_tizen 6.0
  */
@@ -164,7 +168,7 @@ typedef struct _inference_engine_tensor_info {
     std::vector<int> shape; /**< a tensor shape. */
     inference_tensor_shape_type_e shape_type; /**< a tensor shape of the layer. */
     inference_tensor_data_type_e data_type; /**< a tensor type of the layer. */
-    size_t size; /** tensor buffer size. */
+    size_t size; /** tensor buffer size. The size should be height * width * channel count */
     // TODO.
 } inference_engine_tensor_info;