Nexa AI recently unveiled its new OmniAudio-2.6B audio language model, designed to meet the efficient deployment demands of edge devices. Unlike traditional architectures that separate automatic speech recognition (ASR) and language models, OmniAudio-2.6B integrates Gemma-2-2b, Whisper Turbo, and a custom projector into a unified framework. This design eliminates the inefficiencies and delays associated with linking various components in traditional systems, making it especially suitable for resource-constrained computing.

Beijing Baichuan Intelligent Technology Co., Ltd. recently announced the launch of its one-stop solution aimed at helping enterprises achieve private deployment of large models at a low cost while improving efficiency. This solution includes high-quality general training data across the entire link, two models: Baichuan4-Turbo and Baichuan4-Air, as well as a comprehensive toolchain for domain enhancement, meeting the needs of enterprises in various scenarios.

Today, OPPO announced that it has successfully applied the Expert Mixture of Experts (MoE) architecture to mobile devices for the first time globally. This groundbreaking advancement significantly enhances AI processing efficiency, opening new pathways for the further development of terminal AI technologies and laying the foundation for the deep integration of AI and mobile hardware. As AI technology rapidly advances, an increasing number of AI tasks need to be completed on terminal devices. However, large AI models often require powerful computing capabilities, which poses a huge challenge for mobile devices with limited hardware resources. This breakthrough aims to overcome this limitation.

Nvidia has once again demonstrated its ambition in the field of artificial intelligence. The GPU giant has acquired the Seattle-based startup OctoAI for $250 million, marking Nvidia's fifth acquisition of 2024 and a strategic move to solidify its leadership in the AI infrastructure sector. Founded in 2019 and originating from the Apache TVM project at the University of Washington, OctoAI focuses on optimizing AI model performance. Under the leadership of CEO Luis Ceze, the company has successfully transformed AI...

Tencent Cloud announced a strategic partnership with the S.M.A.R.T Entrepreneurs Association in Singapore, aimed at accelerating the deployment of smart digital humans in the Southeast Asian market. The S.M.A.R.T Entrepreneurs Association focuses on assisting global entrepreneurs and senior management teams in building global intelligent enterprises, achieving business transformation and digital transition, and promoting mutually beneficial international business cooperation.

Shen Dou, Executive Vice President of Baidu Group, announced the upgrade of the Baike Computing Platform 4.0 at the 2024 Baidu Cloud Intelligence Conference held today. This new version supports multi-core hybrid training and multi-core adaptation, achieving over 99.5% effective training duration on a ten-thousand-card cluster, significantly improving computing power usage efficiency. In the context of current computing power shortages, the upgrade of Baike 4.0 will help enterprises utilize computing resources more effectively and reduce operational costs. The focus of the upgrade is to enhance the capability of 'multi-core hybrid training', enabling it to achieve 9.

London-based AI startup Rabbitt.AI has secured $2.1 million in funding from TC Group to develop customized LLMs, fine-tune RAG, and create advanced AI applications for deployment. Its platform offers MLOps suites, voice robot AI agents, and emphasizes privacy strategies in AI deployment. Founded by IIT-D alumni, Rabbitt.AI focuses on small, customized, and industry-specific AI models, collaborating with major tech companies like NVIDIA, Meta, and Microsoft. The company has a significant presence in India.

OpenAI announced today that its new flagship model, GPT-4o mini, has been launched on Azure AI, supporting text processing capabilities, with plans to introduce image, audio, and video functionalities in the future.GPT-4o mini is significantly smarter than GPT-3.5Turbo—scoring 82% on the Massive Multitask Language Understanding (MMLU) benchmark compared to GPT-3.5Turbo's 70%—and is over 60% cheaper. This model offers an expanded 128K context window and integrates the enhanced multilingual capa

In the field of artificial intelligence, large language models (LLMs) such as GPT-3 and Llama-2 have made significant strides, capable of accurately understanding and generating human language. However, the vast number of parameters in these models poses challenges in terms of the significant computational resources required for both training and deployment, particularly in resource-constrained environments.Paper Access: https://arxiv.org/html/2406.10260v1Traditionally, to achieve a balance betw

Chipsight AI has announced that the GLM-4-9B and CodeGeeX4-ALL-9B models are now supported for deployment via Ollama. The GLM-4-9B is an open-source pre-trained model released by Chipsight AI, which belongs to the GLM-4 series. It demonstrates outstanding capabilities in semantics, mathematics, reasoning, coding, and knowledge in multiple aspects. The CodeGeeX4-ALL-9B is a multilingual code generation model trained based on the GLM-4-9B, which further enhances the code generation capabilities.Ol

32% of organizations are using generative artificial intelligence in enterprise applications. 60% of organizations are most concerned about the security issues of generative AI. 58% of organizations mainly adopt OpenAI's GPT-4 large language model. Survey data shows that 75% of organizations are currently using large language models.

MobileLLM is a new type of large language model introduced by the Meta AI team, providing solutions for LLM deployment on mobile devices. With a deep and narrow architecture design and parameter optimization, MobileLLM enhances performance. The research opens up new possibilities, making the application of LLM in resource-constrained environments more viable. MobileLLM sets a new standard for LLM deployment on mobile devices, significantly improving performance and broadening new horizons.

Lenovo is working closely with Nvidia to provide AI professional services that help customers quickly deploy hybrid AI solutions. The collaboration aims to strengthen capabilities in the AI field and offer better solutions to customers.

Magic搭 Community collaborates with vLLM and FastChat to provide efficient LLM inference and deployment services. Developers can use vLLM as the inference engine in FastChat, offering high-throughput model inference. FastChat is an open platform for training, serving, and evaluating LLM-based ChatBots. vLLM is an LLM service developed by researchers from the University of California, Berkeley, Stanford University, and the University of California, San Diego.

Microsoft Azure has recently launched GPT-RAG, an ultra-intelligent solution designed to facilitate the operation of large language models in enterprises. The demand for LLMs has surged due to their ability to understand and generate human-like text. GPT-RAG features a robust security framework that adheres to zero trust principles, ensuring that sensitive data is handled with care, thereby providing unparalleled security for enterprises. Additionally, GPT-RAG's architecture includes three key components: data ingestion, a coordinator, and a front-end application.

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.

UK startup HarmonicSecurity successfully raised $7 million in funding, aimed at ensuring the security of enterprise generative AI deployments. Harmonic will develop software to address the issue of unregulated AI applications collecting large-scale corporate data. The investment was led by Ten Eleven Ventures, as startups increasingly focus on the security of generative AI deployments.