From NVIDIA: 2024-03-20 09:00:02

NVIDIA’s RTX AI platform provides tools for Windows developers to create cutting-edge AI features for RTX PCs. GTC showcased new capabilities and models for generative AI, including Gen AI contest winners. Chat with RTX, powered by TensorRT-LLM, allows users to connect local data to AI models for quick, relevant answers.

ChatRTX is expanding to support voice, images, and new models, integrating OpenAI’s CLIP and Google’s Gemma. The Generative AI contest winners developed apps like Rocket League BotChat and OutlookLLM, showcasing the power of NVIDIA’s technology stack. Developers can start with the developer reference project on GitHub.

At GTC, three AI Decoded talks will feature ChatRTX, possibilities for generative AI on RTX, and the impact of AI on creators. Artists like Refik Anadol use NVIDIA technology for innovative installations, while top app developers integrate RTX AI acceleration for faster content creation. The sessions will explore new workflows and ideas for creators.

TensorRT-LLM is integrated with developer frameworks like LangChain and LlamaIndex, allowing easy access to performance benefits for generative AI apps. NIM microservices provide pre-built AI containers with industry-standard APIs for quick deployment. NVIDIA AI Workbench is now generally available, helping developers customize pretrained AI models on RTX GPUs easily.

Developer contest winners showcased apps like Rocket League BotChat and CLARA, which enhance gaming experiences and productivity for users. These apps, powered by NVIDIA’s technology stack, demonstrate the potential of generative AI for personal and professional use. NVIDIA’s ecosystem integrations and tools are driving the development of new Windows apps and features.



Read more at NVIDIA: AI Decoded at GTC: Developer Tools and Apps Accelerating AI