Metric evaluator for Automatic Speech Recognition using the HATS dataset

Speechbrain system

SemDist metric using sentence embeddings

Framework to automatically evaluate transcriptions from an Automatic Speech Recognition system.

Plot a Gaussian Mixture Model

Implementation of a paradigm to make interpretable metrics in ASR.

Minimum Edit Distance experiments

Training parameters for Speechbrain systems

Generate an html file to compare ASR output

Indicate top errors in different systems