AI News

Don't miss any moment of global AI innovation

AI Daily

Daily three-minute AI industry trends

AI Timeline

AI industry milestones

Al Hardware

Lists all AI hardware products.

AI Monetization Guide

Latest Cases

AI monetization case sharing

Image Collection

AI image creation monetization cases

Video Collection

AI video creation monetization cases

Audio Collection

AI audio creation monetization cases

Content Collection

AI content writing monetization cases

AI Tutorials

Latest Tutorials

Free sharing of the latest AI tutorials

AI Product Rankings

AI Product Ranking

Shows total visits ranking of AI websites

AI Traffic Growth Ranking

Track fastest growing AI websites by traffic

AI Traffic Decline Ranking

Focus on AI websites with significant traffic drops

AI Weekly Ranking

Shows weekly visits ranking of AI websites

Popular Country Rankings

United States

AI websites most popular with US users

China

AI websites most popular with Chinese users

India

AI websites most popular with Indian users

Brazil

AI websites most popular with Brazilian users

Popular Category Rankings

Image Generation

Total visits ranking of AI image generation websites

Personal Assistant

Total visits ranking of AI personal assistant websites

Character Generation

Total visits ranking of AI character generation websites

Video Generation

Total visits ranking of AI video generation websites

Popular Open Source Data Rankings

AI Project Ranking

GitHub popular AI projects by total stars

AI Project Growth Ranking

GitHub popular AI projects by growth rate

AI Developer Ranking

GitHub popular AI developer ranking

AI Organization Ranking

GitHub popular AI organization ranking

Popular Open Source Categories

Deepseek

GitHub popular deepseek open source projects

TTS

GitHub popular TTS open source projects

LLM

GitHub popular LLM open source projects

ChatGPT

GitHub popular ChatGPT open source projects

AI Open Source Project Library

Overview

Overview of GitHub popular AI open source projects

Product Library Tool Navigation

Microsoft Azure AI Studio Adds Llama 2 and GPT-4 Turbo with Vision

站长之家

Published inAI News · 1 min read · Dec 15, 2023

Translated Data: Microsoft is introducing the open-source AI model Llama 2 into Azure AI Studio, offering more AI options. Additionally, Microsoft has also launched the OpenAI's GPT-4 Turbo with Vision model. This move strengthens Microsoft's competitive advantage in the AI field.

Model as a Service Llama 2 GPT-4 Turbo with Vision

This article is from AIbase Daily

Welcome to the [AI Daily] column! This is your daily guide to exploring the world of artificial intelligence. Every day, we present you with hot topics in the AI field, focusing on developers, helping you understand technical trends, and learning about innovative AI product applications.

—— Created by the AIbase Daily Team

AI News Recommendations

VideoLLaMA 2: Real-time Video Content Recognition and Interpretation Based on Instructions

With the advancement of artificial intelligence technology, video understanding has become increasingly important. Against this backdrop, the VideoLLaMA 2 project has emerged, aiming to advance video large language model spatiotemporal modeling and audio comprehension capabilities. This project is an advanced multimodal language model that can help users better understand video content.

Jun 14, 2024

4.7k

Google Gemma: Over 7 Billion Parameters with Unique Design Principles

Google has launched a new series of open-source models called 'Gemma'. A comparison between Gemma, Llama 2, and Mistral. The design principles and features of Gemma. Characteristics of the GeGLU activation function used in Gemma.

Feb 28, 2024

450

Llama 2 Defeats GPT-4! Meta Enables Large Models to Self-Reward and Iteratively Improve, Reinforcing the Finality of Synthetic Data for LLMs

AI training AI is set to become a major trend. The teams at Meta and NYU proposed a method for large models to 'self-reward'. This allowed Llama2 to decisively defeat leading models such as GPT-40613, Claude2, and GeminiPro. Llama2-70B stunned the entire AI community by overcoming GPT-4 overnight! Meta's work takes a further step compared to Google's RLAIF paper released last September by employing an ever-evolving reward model for continuous iteration in training the model.

Jan 22, 2024

600

Microsoft Releases Small Language AI Model Phi-2 with Performance Superior to Llama 2 and Mistral 7B

Microsoft Research has released the small language AI model Phi-2, which outperforms Llama 2 and Mistral 7B. Phi-2 has 2.7 billion parameters and can run on laptops or mobile devices. Compared to Google's new Gemini Nano 2 model, Phi-2 shows superior performance, with less toxicity and bias. This provides Microsoft an opportunity to poke fun at Google's Gemini model, as Phi-2 can accurately answer questions and correct students. However, Phi-2 is currently limited to research purposes.

Dec 13, 2023

520

Buddy Compiler Successfully Enables End-to-End Inference of LLaMA 2

Buddy Compiler combines MLIR and PyTorch to achieve end-to-end inference for LLaMA. It uses TorchDynamo as a tracing tool and interfaces with Aten IR to convert AI models into MLIR. The MLIR Python Bindings allow Dynamo Compiler to generate TOSA/LinalgOps. The design principles of Buddy Compiler include standardizing the technical approach, lowering the entry barrier, and maximizing optimization potential.

Nov 23, 2023

640

WhatsApp Launches Llama 2 Powered AI Chatbot

WhatsApp is testing the launch of an AI chatbot powered by Llama 2. Meta has developed a smooth and realistic conversational experience using the advanced Llama 2 large language model. Users can initiate the integrated AI chatbot via shortcuts to perform tasks, providing a more convenient communication experience. Currently, the feature is only available to a limited number of test users, with a full rollout expected in the coming days.

Nov 21, 2023

1.0k

NVIDIA Releases H200: AI Chip Performance Soars by 90%, Llama 2 Inference Speed Doubles

Jensen Huang unveiled the H200 late at night, with performance improvements of 60% to 90%, compatible with the H100. The H200 is based on the Hopper architecture, featuring 141GB of memory and a bandwidth of 4.8TB/s. The inference speed for Llama 2 has nearly doubled, significantly improving the handling of large language models. NVIDIA plans to release new products annually, with H100 production expected to double in 2024 and the introduction of the B100. Global AI companies are stockpiling inventory, and the H200 is set to launch in the second quarter of next year.

Nov 14, 2023

710

Microsoft Research Enables Large Models to Forget Harry Potter: The Memory Erasure Technique of Llama 2

Microsoft researchers have successfully conducted studies on selective forgetting in large models, allowing models to forget specific information rather than the entire content through fine-tuning and reinforcement learning. The research method has certain limitations and requires further validation.

Nov 7, 2023

800

Dell and Meta Collaborate to Promote Local Deployment of Llama 2 Large Language Model

Meta has introduced the next-generation large language model Llama 2, which surpasses GPT-3.5 Turbo and Claude 2 on certain tasks. Released in July 2022, Llama 2 offers three different model sizes: 7 billion, 13 billion, and 70 billion parameters. Meta is collaborating with Dell to assist enterprise customers in deploying Llama 2 models locally instead of solely relying on cloud services. The Llama 2 model is open-source and available for research and some commercial use. Meta has partnered with Microsoft and Amazon.

Nov 1, 2023

790

Meta Releases Llama 2-Long Model, Reducing Computational Demand for Long Text by 40%

Meta's latest Llama2-Long model revolutionizes the handling of long texts without increasing computational demand, delivering exceptional performance. Innovative strategies include continuous pre-training, improved positional encoding, and data mixing, without relying on additional long text data. Llama2-Long excels in both long and short tasks, outpacing other models and holds the potential to transform the field of natural language processing. The model's instruction fine-tuning method is optimized, leveraging a diverse dataset of short prompts, yielding outstanding results, particularly in long-context tasks. Meta's Llama 2-Long model.

Oct 11, 2023

650