executorch_dart 0.6.0

ExecuTorch on-device ML inference for pure Dart using dart:ffi — vision models plus experimental streaming LLM. Android, iOS, macOS, Linux, Windows.

example/README.md

executorch_dart example

Runs an ExecuTorch model from a plain Dart program. No Flutter.

dart run bin/infer.dart /path/to/mobilenet_v3_xnnpack.pte

Compile it to a self-contained bundle, native library included:

dart build cli
./build/cli/*/bundle/bin/infer /path/to/model.pte

Copying this example out of the repo? pubspec.yaml pins resolution: workspace, which only resolves inside this repo's pub workspace. Delete that line first — the existing executorch_dart: ^0.6.0 dependency then resolves normally from pub.dev.