$ show QwenLM/Qwen3-ASR-Toolkit
This toolkit provides a command-line interface for transcribing long audio and video files using the Qwen-ASR API. It overcomes the API's 3-minute duration limit by splitting audio into chunks based on voice activity detection (VAD) and processing them in parallel. The toolkit includes reusable components for audio loading, VAD-based segmentation, and API interaction with retry logic and text post-processing. The main workflow involves loading media, applying VAD to detect speech segments, splitting audio into chunks respecting a target duration, transcribing chunks in parallel via the API, and combining results with optional SRT subtitle generation.
The LLM identified 2 files worth highlighting for reuse:
Provides functions to load audio from local or remote sources (using librosa with ffmpeg fallback), perform voice activity detection (VAD) using the Silero model to split audio into meaningful chunks, and save audio segments. This module is self-contained and can be reused in any project requiring audio preprocessing and VAD-based segmentation.
Implements a wrapper for the Qwen-ASR API with automatic retry logic, file size handling (converting large files to MP3), and post-processing of transcription text to remove character and pattern repetitions. This component is isolated and can be easily integrated into other applications that require robust interaction with the Qwen-ASR service.