[Gneuralnetwork] Patch: dynamically allocate training and output arrays
From:
John Salatas
Subject:
[Gneuralnetwork] Patch: dynamically allocate training and output arrays
Date:
Sat, 19 Mar 2016 13:40:56 -0700
Here is another easy to implement patch.
Instead of defining MAX_TRAINING_POINTS and MAX_NUM_POINTS for input and output arrays respectively, these arrays are now defined as pointers and are initialized in the parser (using malloc).
The same logic can be applied to the other constants defined (MAX_IN, MAX_NUM_NEURONS and MAX_NUM_LAYERS) but it isn't so straightforward to implement.
The patch includes also the patch I previously sent, which BTW was wrong :)