🇧🇷 Versão Português Brasileiro
ASRBench is a framework developed in Python for creating and running benchmarks of audio transcription systems. It allows researchers and developers to compare different transcription systems in terms of accuracy, performance and resource utilization.
The ASRBench organization brings together projects that facilitate the evaluation of transcription systems, including an extensible framework and a command-line interface (CLI) for simple and efficient benchmark execution.
- Simplified configuration: Everything works from a single YAML configuration file, making the process intuitive and agile.
- Extensible framework: Customize benchmarks to meet your specific needs.
- CLI: Run benchmarks quickly with pre-configured systems.
- Comprehensive comparison: Evaluate different systems with detailed metrics.
The core of ASRBench, responsible for providing the structure for configuring and running customized benchmarks.
A command line interface that facilitates the execution of benchmarks with simple configurations and support for already implemented transcription systems.