The XOR truth table is shown in the image below. An XOR gate is replicated and trained with an ANN.
įor this tutorial we will use examples that are already available in the AIfES library. To follow up the examples you have to download and install AIfES (search for aifes ) with the Arduino library manager. This is useful for self-learning systems.įor more information about AIfES, visit. In addition, the ANN structure can also be developed on the PC to insert it later in the Arduino IDE. A porting is therefore possible without any problems. This has the advantage that large networks can be pre-trained on the PC and the source code corresponds to the Arduino implementation. The compiler / hardware must be GCC compatible. The AIfES for Arduino library source code is also executable on the PC. Hardware specific accelerators like Arm CMSIS must be enabled in aifes.h. For this reason AIfES should also run on almost any Arduino or Arduino compatible board. Since AIfES was programmed in pure C, the source code of the AIfES for Arduino library can be used on almost any hardware.
How to train an ANN directly on an Arduino board with AIfES we have already shown in the last tutorial. In this tutorial we use the free and open-source C/C++ Code::Blocks IDE. But we were also frequently asked if it is possible to use AIfES on the PC to develop artificial neural networks (ANN) more easily or to pre-train larger networks to run them later on your Arduino board. For the new versions there will be a new tutorial soon.ĪIfES is compatible with nearly every ML framework, like TensorFlow, Keras or PyTorch. This tutorial is only for AIfES version 2.0.0.
0 kommentar(er)
