Note

Sentis is now called Inference Engine. The documentation has moved to https://docs.unity3d.com/Packages/com.unity.ai.inference@latest. Refer to the new location for the latest updates and guidance. Make sure to update your bookmarks and references accordingly.

Run an imported model

Use Sentis to run an imported model with input data and get the output data.

Page	Description
How Sentis runs a model	Understand how Sentis runs a model.
Create input for a model	Create input data for a model from an array or a texture.
Convert a texture to a tensor	Convert a texture to a tensor, and override texture shape and layout.
Create an engine to run a model	Create a worker, which is the engine that breaks the model down into executable tasks.
Run a model	Run a model in a single frame.
Split inference over multiple frames	Run a model a layer at a time.
Use a command buffer	Use a command buffer to create a queue of Sentis commands, then run the commands on the GPU later.
Get output from a model	Get and log the output data from a model.
Read output asynchronously	Read the output data from a model asynchronously.
Use output data	Convert the output data to an array or a texture, copy the data to the screen.
Manage memory	Call Dispose on workers and tensors.