Open AI's 70+ ChatGPT Plugin Upgrade
Plus: Introducing InstructBLIP. Hugging Face introduces RWKV
Hello, Engineering Leaders and AI Enthusiasts,
Welcome to the 19th edition of The AI Edge newsletter. In today’s edition, we bring you the latest revolutionary updates from Open AI for ChatGPT Plus users. Thank you everyone who is reading this. 😊
In today’s edition:
🌐ChatGPT, now with internet access & 70+ Plugins
🧠InstructBLIP: the new vision-language instruction-tuning framework
🚀Hugging Face introduces integration of a new architecture: RWKV
📚 Knowledge Retrieval Architecture for LLMs (2023)
Let’s go!
ChatGPT Plus, with superpowers of web browsing & 70+ Plugins
The latest release notes of ChatGPT highlight various updates and new features added to the model, such as the ability to turn off chat history, export data, and use AI plugins. In addition, there are updates on the model's performance, factuality, and mathematical capabilities.
ChatGPT Plus subscribers get early access to experimental features, such as the ability to choose between different versions of ChatGPT and new experimental models that can browse the internet and use third-party services.
Why does this matter?
With a whole bunch of AI announcements by Google last week, Microsoft and OpenAI responded by releasing the ChatGPT plugins for everyone. However, amid the constant battle between these tech behemoths, users like you and us are getting the most benefit. The plugins can be extremely useful in a variety of cases, especially the ‘browser’ access.
The Multitasker: How InstructBLIP handles both Vision and Language
InstructBLIP is a vision-language model that utilizes instruction tuning and instruction-aware visual feature extraction to improve performance on vision-language tasks.
It is based on pre-trained BLIP-2 models and achieves state-of-the-art performance on multiple benchmarks, including zero-shot evaluation. The model is designed to effectively integrate visual and linguistic information to solve complex language-domain tasks. Plus, all InstructBLIP models have been open-sourced
Why does this matter?
InstructBLIP has potential applications in a wide range of fields where vision and language are important, such as Healthcare, where InstructBLIP can analyze medical images and provide relevant information or diagnoses, and Robotics, where InstructBLIP can analyze visual data from robots and provide relevant instructions or feedback.
Hugging Face introduces RWKV, combining the best of RNNs & transformers
Hugging Face introduced a new architecture, RWKV, which takes the best from transformer-based models and RNNs (Recurrent Neural Networks) to provide faster inference and theoretical infinite context length.
It has been integrated into the Hugging Face transformers library and comes with the addition of the "chat" version of the model, called RWKV-raven. Moreover, the GPUs for training the RWKV models are donated by Stability AI.
Why does this matter?
Integrating a blend of transformers and RNNs is a significant move from Hugging Face as it expands the range of available architecture options for developers/researchers working on NLP tasks. It also offers a powerful and flexible architecture, expanding the possibilities in NLP with an advanced approach to generative modeling.
Can’t wait to see how such novel ideas push the boundaries of AI or help build upon them!
Knowledge Nugget: Knowledge Retrieval Architecture for LLMs
In most practical applications of LLMs, we want them to generate specific answers rather than generic ones based on the model’s training set, don’t we?
This insightful article discusses a solution to overcome the shortcomings of knowledge retrieval methods of existing LLMs and generate more contextual, accurate responses.
For an advanced approach to the solution, the author discusses several research developments that can enter into the solution architecture to enhance the LLMs for better knowledge retrieval.
Why does this matter?
New advancements associated with LLMs are announced every day. However, the insights from the above article enable businesses, AI developers, and researchers to design more effective, accurate, and context-aware LLM applications.
What Else Is Happening
🎬 AI's Stable Diffusion in action and Coca-cola in a frame! Watch this Ad (Link)
🤖 Amazon developing a secret home robot with AI - could rival Rosie (Link)
💡 Microsoft AI research introduces Automatic Prompt Optimization framework for LLM prompts (Link)
🎮 Run Llama 13B with just a 6GB graphics card (Link)
💻 Google unveils new AI Supercomputer powered by Nvidia H100 GPUs (Link)
🚀 Boost.ai introduces LLM-enriched features in Version 12 of conversational AI platform (Link)
Trending Tools
RomptAI: Fine-tune AI products with extensive A/B testing on prompts. Uncover top performers with unbiased ratings.
Databerry: Build ChatGPT Plugins to connect custom data to ChatGPT. Get insights and statistics on user interactions.
WPTurbo: Get WordPress code snippets with ease. AI and online generators help generate clean and optimized code.
AI Carousels: Simplifies crafting visually engaging carousels for social media. Boost online presence without design expertise.
Ai2Page: Generate relevant one-page websites. Chat for updates, regenerate blocks, edit texts and images, and publish.
WebsiteColorsAI: Uses AI to discover and analyze color palettes used on top websites. Saves time and optimizes design workflow.
ExpiredAI: Spot valuable recently expired domains and catch them at the right time.
Imgartfis Creators: Chrome extension to easily create AI images directly from a Google search.
That's all for now!
If you are new to ‘The AI Edge’ newsletter. Subscribe to receive the ‘Ultimate AI tools and ChatGPT Prompt guide’ specifically designed for Engineering Leaders and AI enthusiasts.
Thanks for reading, and see you tomorrow.