package com.codename1.ai.inference
Reusable on-device inference sessions for .tflite models.
Android uses LiteRT. iOS uses TensorFlow Lite Objective-C and may select
the Core ML delegate. The builder links the runtime only when an application
references InferenceSession. Other
targets expose the same API with an explicit unsupported fallback.
Types
class InferenceException | Reports model validation, delegate selection, allocation, or invocation failures from the portable inference API. |
class InferenceOptions | Configures how a reusable InferenceSession is created. |
class InferenceSession | Reusable, native on-device session for a TensorFlow Lite model. |
class ModelCache | Downloads a large model into app-private storage and exposes it as a file-backed ModelSource. |
class ModelSource | Describes where an InferenceSession should obtain a .tflite model. |
class Tensor | Immutable named tensor passed to or returned from an inference session. |
class TensorInfo | Immutable metadata for one model input or output. |
enum TensorType | Tensor element types supported by the portable LiteRT contract. |