AsrBench

Evaluate, compare and find the best model for audio transcription.

ASRBench is a framework developed in Python to create and run benchmarks of audio transcription systems. It allows researchers and developers to compare different transcription systems in terms of accuracy, performance and resource utilization.

ASRBench also offers a Command Line Interface (CLI) with the main audio transcription systems already implemented and ready to use. With the CLI, you can run audio transcription benchmarks simply by writing the configuration file. However, if you wish, you can add new transcription systems according to your needs.