The most recent update for the week of April 17th, 2023, contains the most recent developments and announcements in the fields of data science, machine learning, and artificial intelligence. This week, the spotlight is on leading actors such as Hugging Face, AWS, Microsoft, and Stability AI who have introduced revolutionary innovations.
HuggingGPT: Using LLMs to Solve Difficult AI Tasks with Hugging Face Models
Recent research from Zhejiang University and Microsoft Research Asia investigates the use of large language models (LLMs) as a controller to administer existing AI models available in communities such as Hugging Face.
AWS Joins the Race for Generative AI with the Bedrock and Titan Foundation Models
Amazon Web Services announced their entrance into the race for generative AI with the release of the Amazon Bedrock and Titan foundation models. Amazon intends to democratize access to generative AI technology, catering to clients in a variety of industries and use cases. This revolutionary innovation positions Amazon as a formidable rival in the swiftly expanding AI market.
AWS also announced the general availability of Amazon EC2 Trn1n instances powered by AWS Trainium and Amazon EC2 Inf2 instances fueled by AWS Inferentia2, the most affordable cloud infrastructure for generative AI.
Microsoft Releases Multimodal Chatbot Visual ChatGPT Source Code
Microsoft Research has recently released Visual ChatGPT, a chatbot system capable of generating and manipulating images in response to textual human commands. The system supports multimodal interactions by combining OpenAI’s ChatGPT with 22 distinct visual foundation models (VFM).
Stability AI Launches Its First Language Model from the StableLM Suite
Stability AI has released StableLM, a novel open-source language model. The Alpha version of the model is available with 3 billion and 7 billion parameters, with subsequent models containing 15 billion to 65 billion parameters. Subject to the terms of the CC BY-SA-4.0 license, developers are free to investigate, utilize, and modify our StableLM base models for commercial or academic purposes.