Can Wyser handle different accents?
For example, UK regional & non-native speakers
We use our large database of UK accent data, captured via our AccentLog project (soon to be available to everyone who wants to submit their accent), to benchmark Large Language Models (LLMs). ASSIST is LLM-agnostic which means that if we determine a new model is better than the one we currently use, we can move to the new model. This ensures that we’re always using best in class technology.