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.
We cover the state of the art Python Speech Recognition technologies. There are three open source libraries covered, wav2letter, SpeechRecognition, and DeepSpeech. Then we cover AssemblyAI's speech to text API as a super simple cloud solution that also offers custom vocabulary, speaker diarization, and paragraph extraction.
On Tuesday, August 31st, AWS had an outage in their us-west-2 region. At 18:00 UTC that day, we experienced an increase in 5xx error codes returned by our API, as well as a slowdown in transcription turnaround time. The AWS outage impacted a single AWS availability zone, usw2-az2. We would like to take this opportunity to share our post mortem of this event.
We explore if the frequency of negative news in the popular podcasts The Daily and Up First can predict the stock market. We use Python Selenium with Chromedriver to crawl Listen Notes and send the links to be transcribed with AssemblyAI's speech to text API.
How to use AssemblyAI's mp4 to text API to generate a subtitle file to add to your Mux video programmatically in Python. This speech to text example shows how to get the .srt and .vtt subtitle files from AssemblyAI, upload them to S3, and then use Mux's Python SDK to upload a video with subtitles.
Learn how to use Python to set up Twilio Voicemail. This tutorial shows how to use a set of Python scripts to programmatically access Twilio's API and set up a voicemail system that can record your calls and which you can download .mp3 recordings from.
What is a burner phone number? It's one you can drop for things you don't want to give out your real phone number to and can get rid of easily. This tutorial shows you how to build a Python burner phone with Twilio and AssemblyAI that will transcribe your voicemails.
A tutorial on how to use the components of the Python Click library to intuitively and easily build simple to complex command line interface (CLI) applications. This tutorial covers styling, passing context, creating your own pass decorators, nested commands, and how to use multiple command groups.
An easy to follow Python tutorial on how to do speech recognition in under 25 lines of code. We extend our short example using AssemblyAI's API to a longer script and build a Python project that can automatically do speech recognition and transcribe an mp3 into a txt file.
Since being first developed and released in the Attention Is All You Need paper Transformers have completely redefined the field of Natural Language Processing. In this blog, we show you how to quickly fine-tune Transformers for numerous downstream tasks, that often perform really well out of the box.
In this post, we’re going to show you how to transcribe your Zoom recordings by connecting Zoom’s API with AssemblyAI’s automatic speech recognition API. In just a few lines of code, you'll see how you can accurately transcribe your Zoom recordings!
We’re excited to announce that we’ve open sourced another project! `drone-deploy-ecs` is a Drone plugin that enables you to deploy updates to ECS. Our engineering team has recently made the decision to migrate from Docker on EC2 to AWS ECS. We knew that moving to ECS would require us to refactor our deployment processes, so we figured we’d wrap our deployment process into a single tool that fit into our CICD solution.
At AssemblyAI, we use Drone as our primary CI/CD tool. It's dead simple to set up and operate which frees us up to build out our product. Recently we decided to figure out how to build a cost-effective, easily-scalable Drone worker fleet for our GPU instances.
As part of our core research and development efforts to continue pushing the state of the art of speech recognition accuracy, in this post, we explore speech recognition architectures that are gaining new popularity in both academia and industry settings.