🔥 Trending AI Research: Let’s learn something new from the trending papers.

💻 Some Cool Github Repos: Take a deep dive into the world of advanced AI with these trending Github repos

🏃 Let’s Practice Some Generative AI: Let’s do some fun in the lab by trying some prompts and AI tools.

🛎️ Trending Tools: Check out some cool AI tools picked up by our editorial team.

Read Time: 5 Minutes

🔥Trending AI Research

1️⃣ H2O.ai Introduces h2oGPT: A Suite of Open-Source Code Repositories for Democratizing Large Language Models (LLMs)

With the aim of promoting a more open and accessible AI ecosystem, h2oGPT is a testament to H20.ai’s commitment to democratizing AI. This project’s main objective is to create the finest truly open-source substitute for closed-source methods. In partnership with the open-source community, H2O.ai has made a number of improved h2oGPT models available under completely permissive Apache 2.0 licenses. These models have different parameters ranging from 7 to 40 billion and are prepared for commercial use, enabling unrestricted use by people, groups, and businesses in their applications.

2️⃣ Microsoft Research Introduces LongNet: A Transformer Variant That Can Scale Sequence Length To More Than 1 Billion Tokens With No Loss In Shorter Sequences 

The main design tenet is that attention allocation falls off rapidly as the distance between tokens increases. They demonstrate that it achieves a logarithm dependence between tokens and linear processing complexity. This addresses the conflict between the accessibility of all tokens and the finite amount of attention resources. LONG NET may be converted into a dense Transformer in the implementation, which supports the standard Transformer optimizations (such as kernel fusion, quantization, and distributed training) without problems. By using linear complexity, LONGNET can divide the training among nodes and overcome the limitations of both CPU and memory. Since the vanilla Transformer suffers from quadratic complexity, this enables us to efficiently scale up the sequence length to 1B tokens with practically constant runtime.

3️⃣ Salesforce AI Research Releases CodeGen2.5- a small but mighty LLM for code 

CodeGen2.5 with 7B is on par with >15B code-generation models (CodeGen1-16B, CodeGen2-16B, StarCoder-15B), less than half the size. It is trained on trained on 1.5T tokens. CodeGen2.5 features robust infill sampling and fast sampling optimized for serving with Flash attention, both of which enable the productization of these models for coding assistants. In the future, we will push the boundaries of such models further.

4️⃣ Princeton Researchers Introduce InterCode: A Revolutionary Lightweight Framework Streamlining Language Model Interaction for Human-Like Language-to-Code Generation

In recent times, the development of models that can automatically produce code from natural language specifications has gained popularity. Though these models have demonstrated impressive performance on static benchmarks due to the extensive pre-training over thousands of codebases, there are also certain limitations, such as typos, gaps between the process of creating the code and its execution, limited human involvement, and so on.

To address these challenges, researchers from the Department of Computer Science at Princeton University have proposed a lightweight and flexible framework called InterCode that facilitates interactive coding as a standard reinforcement learning (RL) environment. In InterCode, code is treated as actions, and execution feedback is considered as observations. This RL-based method makes coding more iterative and can be used with many programming languages and environments because it is made to be language and platform-independent.

💻 Github Repos

➡️ hiyouga / ChatGLM-Efficient-Tuning: Fine-tuning ChatGLM-6B model with PEFT.

➡️ chat2db / Chat2DB: An intelligent and versatile general-purpose SQL client and reporting tool for databases that integrates ChatGPT capabilities.

➡️ microsoft / unilm: Unified pre-training for language understanding and generation.

➡️ vllm-project / vllm: vLLM is a fast and easy-to-use library for LLM inference and serving.

➡️ databrickslabs / pyspark-ai: English SDK for Apache Spark.

🏃Let’s Practice Some Generative AI

Storybird.ai is an innovative tool that harnesses the power of AI to enable users to create captivating visual stories effortlessly. It addresses the challenge of crafting engaging narratives and illustrations by providing an intuitive platform powered by advanced language models and generative techniques.

We have created a video to explain How to use Storybird.ai

🛎️ Trending Tools

Notion: Notion is aiming to increase its user base through the utilization of its advanced AI technology. Their latest feature, Notion AI, is a robust generative AI tool that assists users with tasks like note summarization, identifying action items in meetings, and creating and modifying text.

AdCreative AI: Generate conversion-focused ad creatives and social media post creatives in a matter of seconds using Artificial Intelligence

Taplio: Transform your LinkedIn presence with Taplio's AI-powered platform. Spend just 10 minutes a day to elevate your personal brand.

Booth AI: Experience the extraordinary with Booth.ai: AI-generated photos that redefine product photography. Effortlessly create pro-quality lifestyle shots by specifying your vision and uploading sample product images.

Sheet+: Unleash Sheet+, the formula genius! This AI-powered tool speeds up Google Sheets and Excel, transforming text into precise formulas in a flash.

Slayer AI: Slayer AI revolutionizes podcasting by crafting tailor-made shows in mere seconds.

🚀 Check Out 700+ AI Tools in AI Tools Club

 Bonus Section

Sponsored Section

🤝 Partner with us