IRecognitionTrainingModel<T>
Interfaces · AiDotNet.Interfaces
A text recogniser that trains on transcriptions, using its paper’s objective (CTC for CRNN, teacher-forced cross-entropy for TrOCR).
Known implementations: OCRBase<T>
Methods
| Method | Summary |
|---|---|
void TrainRecognition(Tensor<T> images, IReadOnlyList<string> transcriptions) |
One training step on cropped text-line images and the text each one shows. |