Shypd.aiThe largest open index of AI toolsspeech-to-text-benchmark is an open-source framework for evaluating speech-to-text engines. It provides a minimalist and extensible platform for benchmarking various engines and datasets.
Trending
About
speech-to-text-benchmark is an open-source, minimalist, and extensible framework designed for evaluating the performance of different speech-to-text engines. It allows users to benchmark engines like Amazon Transcribe, Azure Speech-to-Text, Google Speech-to-Text, OpenAI Whisper, and Picovoice Cheetah/Leopard against various datasets including LibriSpeech, TED-LIUM, Common Voice, and VoxPopuli. The framework calculates key metrics such as Word Error Rate (WER), Punctuation Error Rate (PER), Core-Hour for computational efficiency, Word Emission Latency for streaming engines, and Model Size. It supports multiple languages and provides clear instructions for setting up and running benchmarks, making it a valuable tool for researchers and developers in speech recognition.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
