# Timestamp annotation

A Whisper model fine-tuned for the Levantine dialect (Israeli Arabic) to improve the performance of automatic speech recognition for this specific Arabic variant.

Whisper is a Transformer-based encoder-decoder model for speech recognition and translation tasks, supporting multilingual processing.

A sequence-to-sequence model supporting English automatic speech recognition (ASR), capable of outputting normalized text, timestamp annotations, and multi-speaker segmentation.

ONNX weight version of OpenAI Whisper-tiny.en model, designed for Transformers.js, used for English speech transcription.