---
tags:
- automatic-speech-recognition
- asr
- amharic
- speech-to-text
- wav2vec2
- speechbrain
- social-media
---

# Amharic Speech-to-Text Transcription Model

This model transcribes Amharic speech to text. It's built on **Facebook's Wav2Vec2** and trained using **SpeechBrain**.

## Intended Use

Its main purpose is to transcribe audio from **Instagram, YouTube, and TikTok video content** for further analysis (e.g., trend identification, content moderation).

## Limitations

Performance may vary with audio quality, background noise, and informal speech commonly found in social media.