LLaVA-1.6 Released! Surpassing Gemini Pro, Enhanced Inference Performance

No Training Needed! Q-Filters Enable Efficient Compression of KV Cache and Improved Inference Performance

Have You Used DeepSeek R1? Its Inference Performance Is Unmatched, and Meta Is Already Panicking!

NVIDIA GB300/B300 GPU Launches! Dramatic Increase in Inference Performance, Supply Chain Shake-Up

Former Character.AI Founder Returns to Google to Assist in the Development of the Gemini Project

Google's acquisition of the AI startup Character.AI and the return of its former CEO Noam Shazeer as co-technical lead for the Gemini project demonstrate its ambitions in the field of artificial intelligence. Shazeer will collaborate with experts like Jeff Dean and Oriol Vinyals to ensure that the Gemini project maintains an edge in competition with ChatGPT. This acquisition not only strengthens Google's AI research team but could also enhance AI conversational capabilities.

Gemini Pro 1.5 and Its Million Context Feature Now Open to Everyone

Gemini Pro 1.5 and the million context feature are now available to all users, expanding the possibilities of AI applications. The new features include a long context capability for video modalities, allowing AI to handle longer and more complex video content. Gemini Pro 1.5 has demonstrated powerful processing capabilities during testing, leading the development in the AI field. Google provides API support for developers using Gemini 1.5 Pro, encouraging the birth of innovative products. The model architecture and capabilities of Gemini 1.5 Pro make it easier to process lengthy information.

Apple is negotiating to have Google's Gemini provide AI functionality support for the iPhone

Apple Inc. is in internal talks to integrate Google's next-generation generative AI system, Gemini, into iPhone software, which could lead to significant changes in the AI field. Sources say that Apple and Google are actively negotiating the details of licensing Gemini to provide some AI-driven features for this year's new iPhone. Gemini may be integrated into iPhone applications such as Siri and the camera app, significantly enhancing the interactive and creative experience of AI. Apple is also in discussions with OpenAI regarding the use of its generative technology.

Google Bard Global Update: Supports 40 Languages and Adds Image Generation Feature

Google Bard announced a global update today, expanding the Gemini Pro large language model capabilities to 40 languages and adding an image generation feature. This update is a significant step in Bard's development, making it easier to use and more competitive. Gemini Pro is Google's most advanced LLM to date. Bard ranks second in the Chatbots Arena leaderboard.

Draw a Circle to Search: Google Gemini Pro Integrated into Flagship, Starting the Mobile AI War

Samsung grandly launched the latest Galaxy S24 series smartphones in San Jose, California, packed with AI features. Netizens have reported that despite the abundance of AI features in Samsung's new flagship, many functionalities are poorly executed. The Circle to Search feature allows Samsung to 'summon' and draw a circle around interesting items on any interface, and the AI will consolidate all related information. The Galaxy S24 series is the first smartphone equipped with Gemini Pro and Imagen2.

The domestic open-source project SwiftInfer has achieved infinite streaming input inference, successfully improving large model inference performance by 46%

["The domestic open-source project SwiftInfer has achieved infinite streaming input inference, successfully improving large model inference performance by 46%", "SwiftInfer employs techniques such as attention sink mechanism, window attention optimization, and KV Cache optimization to successfully implement infinite streaming input inference, resulting in a 46% increase in inference throughput", "The improvement in inference performance by SwiftInfer released by the Colossal-AI team is evident for large models."]

Does Gemini Pro's Inference Capability Surpass GPT-4? Stanford and Meta's Joint Research Validates It

Gemini Pro excels in common sense reasoning tasks, and research from Stanford and Meta suggests that Gemini is underrated. Researchers designed a new test set and found that Gemini's performance in complex reasoning tasks is significantly stronger than before. The evaluation includes 11 language-based datasets and one multimodal dataset, showing that Gemini Pro is on par with GPT-3.5. While GPT-4Turbo leads in overall performance, Gemini Pro slightly outperforms in specific tasks. Inference experiments demonstrate that Gem