Dylan Fox - AssemblyAI Blog

Releasing our v8 Transcription Model - 18.72% Better Accuracy

Today, we’re happy to release our most accurate Speech Recognition model to date - version 8 (v8). This updated model delivers up to 18.72% better accuracy across many types of audio.

Deep Learning

How to Train Large Deep Learning Models as a Startup

Training large deep learning models is expensive and slow. Yet, startups are all about iterating fast. In this post, we share the lessons we've learned over the past few years.

Deep Learning

Is Word Error Rate Useful?

What is Word Error Rate? Word Error Rate is a measure of how accurate an Automatic Speech Recognition (ASR) system performs. Quite literally, it calculates how many “errors” are in the transcription text produced by an ASR system, compared to a human transcription.

Announcements

A New API Endpoint to Paginate Through Historical Transcripts

We are excited to make available our new List Endpoint! This gives developers the ability to query for, and paginate through, all of their historical transcriptions.

Announcements

Content Safety Detection is now GA!

Automatically transcribe audio and video files, and surface sensitive content, such "Hate Speech" or "NSFW" content, found within the audio.