Unity Sentis

Sentis is a neural network inference library in Unity. You can use Sentis to import trained neural network models into Unity, then run them natively in real time on any platform supported by Unity. You can run the model on GPU or CPU. Using Sentis requires some experience working with machine learning models, for example in frameworks such as TensorFlow or PyTorch.

Demand group:

"Game developers can use Sentis to integrate machine learning functions in Unity games, such as speech recognition, object detection, etc."

Example of usage scenario:

Identify objects in game scenes using a pretrained image classification model

Integrate speech recognition model into Unity chatbot

Use the target detection model to implement the smart camera tracking function in Unity

Product features:

Import ONNX model

Run models on GPU and CPU

Get and set input and output tensors

Analyze model performance using Unity tools