RobotsMali_ASR_DEMO / agents.md
AI Assistant
Fix: revert to robust Docker SDK, add agents.md, and finalize app.py logic
a3e5374

RobotsMali ASR Agent

This agent provides Automatic Speech Recognition (ASR) for Bambara using state-of-the-art models from RobotsMali.

Description

This space implements several models optimized for Bambara language transcription, including:

  • Soloni V3 (TDT-CTC): A hybrid architecture for fast and accurate transcription.
  • Soloba V3 (CTC): A robust CTC-based model.

The agent can take audio input (via upload or internal path) and return a full text transcription.

Tools

transcription

Transcribes an audio file into text.

  • audio: The audio file to transcribe (Path or URL).
  • model_name: The name of the model to use (default: "Soloni V3 (TDT-CTC)").

Returns: String containing the transcription.