Improved Punctuation Restoration & Truecasing Models

Check out our Improved Punctuation Restoration & Truecasing Models, dive into our Speech-to-Text AI models and celebrate with us as we near 100K YouTube subscribers.

Improved Punctuation Restoration & Truecasing Models

Hey ๐Ÿ‘‹, this weekly update contains the latest info on our new product features, tutorials, and our community.

๐Ÿš€New Punctuation Restoration and Truecasing Models

Weโ€™re introducing new models for Punctuation Restoration and Truecasing, which outperform our previous production modelson the following metrics:

  • Enhanced Casing Accuracy: Improvements in the handling of mixed-case words (+39% F1 score), acronyms (+20% F1 score), and capital-case (+11% F1 score).
  • Upper-case Letter Classification: A 17% improvement on average across test datasets for upper-case letter detection.
  • Punctuation Precision: An 11% increase in punctuation accuracy (F1 score).
  • Evaluator Preference: Human evaluators show a 61% preference for the new model compared to the previous one.

Join Us At AWS re:Invent

We'll be in-person at AWS re:Invent from November 27 through the 30th. Find us at The Generative AI Partner Pavilion Booth #372, where we'll have folks from our Engineering, Sales and Developer Relations teams available to chat.

Refreshed Speech-to-Text Documentation

Check out our refreshed AI Models documentation pages that make it easier to build with our new SDKs:

  • Summarization: Distill important information from your audio into a succinct text summary.
  • Sentiment Analysis: Get a detailed analysis of positive, negative, or neutral sentiment conveyed in the audio, along with a confidence score for each result.
  • Auto Chapters: Summarize audio data into chapters over time and make it easy for users to navigate and find specific information.
  • PII Redaction: Automatically remove sensitive information about individuals from your transcript.

๐ŸŽ‰(Almost!) 100K Subscribers on YouTube

We're close to 100K on YouTubeSubscribe now and be part, of our growing community! Here's some of our most-watched videos:

Fresh From Our Blog

Key phrase detection in audio files using Python: Learn how to identify key phrases and important words using Python and AssemblyAI. Read more>>

Automatically determine video sections with AI using Python: Build an application that automatically determines video sections. Learn how to generate section titles with LLMs and how to format the information for YouTube chapters. Read more>>

Automatically summarize audio and video files at scale with AI: Learn how AI summarization helps developers and product teams build exciting features that automatically summarize audio and video data. Read more>>

๐Ÿ”ฎ MemGPT: The Future of LLMs with Unlimited Memory: Explore how MemGPT is changing (LLMs) with virtually unlimited token size and how to run it locally on your machine.

Transcribe Multiple Files Synchronously using Webhooks with AssemblyAI: Learn how to use webhooks with AssemblyAI's Speech-to-text API and build an app demonstrating the use of webhooks together with custom parameters.

LangChain Pandas Agent and GPT-4 for Data Analysis: Learn how to set up and use the Pandas DataFrame Agent with GPT-4 from OpenAI.