Minor indentation fix
Report from https://github.com/nnstreamer/nntrainer/pull/993#discussion_r594884810
**Self evaluation:**
1. Build test: [x]Passed [ ]Failed [ ]Skipped
2. Run test: [x]Passed [ ]Failed [ ]Skipped
Signed-off-by: Parichay Kapoor <pk.kapoor@samsung.com>
The management for the above 3 kinds of memory is done depending on the mode of execution of the model, namely `inference` or `training` mode. The memory management of these tensors is performed by the `Manager` inside `NNTrainer`.
-The memory for the `weights` are allocated at the time of initializing the model, where the `weights` are either initialized using the provided initializer or loaded from the saved model file. The memory allocated for `weights` is freed
-upon `destruction` of the model object containing the `weights`.
+The memory for the `weights` are allocated at the time of initializing the model, where the `weights` are either initialized using the provided initializer or loaded from the saved model file. The memory allocated for `weights` is freed upon `destruction` of the model object containing the `weights`.
The memory for the `Input, Output and Label` tensors and `Variables` is allocated lazily, and de-allocated once its usage is finished.