formwise_ml library
Optional ML-powered validation module for formwise.
Provides on-device TFLite model loading, text preprocessing, and a base class for building ML-backed validators.
Requires the tflite_flutter package and platform-specific setup.
See https://pub.flutter-io.cn/packages/tflite_flutter for native library setup.
import 'package:formwise/formwise_ml.dart';
class MyValidator extends MlValidator {
MyValidator() : super(modelAsset: 'models/my_model.tflite');
// ...
}
Classes
- EmailMlValidator
- ML-powered email typo detector using an on-device TFLite model.
- MlValidationResult
- The result of an ML-powered validation.
- MlValidator
- Base class for ML-powered validators that run on-device TFLite models.
- ModelInfo
- Describes the input/output tensor shapes and types of a loaded model.
- ModelLoader
- Manages loading and caching of TFLite models for on-device inference.
- OutputDecoder
- Decodes model output probabilities into predictions.
- TextPreprocessor
- Converts text to numeric tensors for TFLite model input.