NVIDIA has announced new developments at Computex 2024, introducing advanced RTX technology to enhance AI assistants and digital humans on GeForce RTX AI laptops.
Project G-Assist
NVIDIA showcased Project G-Assist, an RTX-powered AI assistant that offers context-aware assistance for PC games and applications.
Demonstrated with the game ARK: Survival Ascended from Studio Wildcard, Project G-Assist provides game-related insights, answers questions about game elements, and personalizes responses based on the player’s session.
Additionally, it optimizes gaming system performance, adjusts graphics settings, and manages power consumption.
NVIDIA NIM and ACE Digital Human Platform
Additionally, NVIDIA introduced the first PC-based NVIDIA NIM (NVIDIA Inference Microservices) for the ACE digital human platform.
This technology reduces deployment times for developers and delivers high-quality local inference for tasks like natural language understanding and facial animation.
The debut includes a demo of Covert Protocol, featuring NVIDIA Audio2Face and NVIDIA Riva for local speech recognition.
RTX AI PC Laptops
New RTX AI PC laptops from ASUS and MSI, equipped with GeForce RTX 4070 GPUs, offer power-efficient systems running Windows 11 AI. These laptops will receive a free update to Copilot+ PC experiences upon availability.
Windows Copilot Runtime
NVIDIA and Microsoft are collaborating to enhance Windows native and web apps with GPU-accelerated small language models (SLMs).
These models, part of Windows Copilot Runtime, will provide capabilities like content summarization and task automation, using NVIDIA RTX GPUs and other AI accelerators.
A developer preview of the API will be released later this year.
RTX AI Toolkit
NVIDIA introduced the RTX AI Toolkit, a collection of tools and SDKs for customizing, optimizing, and deploying AI models on RTX AI PCs.
Available later this month, the toolkit includes QLoRa tools for model customization and NVIDIA TensorRT for model optimization, resulting in up to 4x faster performance and 3x smaller models.
The NVIDIA AI Inference Manager SDK simplifies ACE deployment on PCs and orchestrates AI inference across devices and the cloud.
AI for Content Creation
NVIDIA is integrating RTX AI acceleration into creative apps. The RTX acceleration will enhance tools like Automatic1111 and ComfyUI for Stable Diffusion, delivering significant performance improvements.
RTX Remix, a platform for remastering classic games, will be made open source, allowing modders to create enhanced versions of games with full ray tracing and other advanced features.
The RTX Remix Toolkit will also be accessible via a REST API and an SDK, expanding its use in various applications and games.
RTX Video
NVIDIA RTX Video, an AI-powered super-resolution feature, is now available as an SDK for developers. This will allow native integration of AI for video upscaling, sharpening, and HDR conversion in applications.
Upcoming integrations include video editing software like Blackmagic Design’s DaVinci Resolve and Wondershare Filmora, enabling editors to enhance video quality up to 4K resolution. The VLC media player will also add RTX Video HDR capabilities.
Furthermore, NVIDIA has announced the general availability of its ACE generative AI microservices, aiming to advance digital human technologies in sectors like customer service, gaming, and healthcare. This launch includes:
- NVIDIA Riva ASR, TTS, and NMT: Technologies for automatic speech recognition, text-to-speech conversion, and translation.
- NVIDIA Nemotron LLM: For language understanding and contextual responses.
- NVIDIA Audio2Face: For realistic facial animation from audio tracks.
- NVIDIA Omniverse RTX: For real-time, path-traced realistic skin and hair rendering.
Newly revealed technologies include:
NVIDIA Audio2Gesture: Generates body gestures from audio, available soon.
NVIDIA Nemotron-3 4.5B: A small language model designed for low-latency, on-device RTX AI PC inference, now in early access.
Digital Humans on 100 Million RTX AI PCs
Previously, ACE was available as NIM microservices for data centers. Now, NVIDIA is extending ACE to 100 million RTX AI PCs and laptops, including Nemotron-3 4.5B SLM, Audio2Face, and Riva ASR on-device models, which will soon enter early access.
The new AI Inference Manager SDK simplifies ACE deployment by configuring necessary AI models and orchestrating inference across PCs and the cloud.
At Computex, NVIDIA showcased an updated Covert Protocol tech demo, developed with Inworld AI. It uses Audio2Face and Riva ASR on GeForce RTX PCs, allowing players to interact with digital-human NPCs using conversational language.
Expanding Digital Human Ecosystem
Developers from companies like Aww Inc., Dell Technologies, Gumption, Hippocratic AI, Inventec, OurPalm, Perfect World Games, Reallusion, ServiceNow, Soulbotix, SoulShell, and UneeQ are leveraging ACE technologies. Highlights include:
- Aww Inc.: Plans to use Audio2Face for real-time animation of its virtual celebrity, Imma.
- Perfect World Games: Using ACE in its tech demo, Legends, for realistic, multilingual AI NPC interactions.
- Inventec: Enhancing its healthcare AI agent in the VRSTATE platform with Audio2Face for better virtual consultations.
- ServiceNow: Showcased ACE NIM in its generative AI service agent demo for improved customer and employee interactions.
- Dell Technologies: Launched the Dell Generative AI Solution for Digital Assistants, enhancing customer engagement through natural conversations.
Celebrating Digital Human Startups at Computex 2024
NVIDIA showcased its generative AI tools by creating a “digital Jensen” avatar, featuring NVIDIA CEO Jensen Huang’s voice and style in multiple languages.
This was achieved using Synthesia, Hour One, and ElevenLabs’ AI speech technology, along with Voicemod’s AI voice technology for the keynote’s ending theme song.
SFF-Ready Enthusiast GeForce Card program
NVIDIA has announced the SFF-Ready Enthusiast GeForce Card program for gamers building smaller systems with GeForce RTX 40 graphics cards.
This program involves collaboration with board partners, case manufacturers, PSU vendors, and other component designers.
Key details:
- GPU Compatibility: Includes RTX 4070, 4070Ti, and 4070 SUPER.
- Size Requirements: Maximum thickness of 2.5 slots (50mm), 304mm in length, and 151mm in height, including a 12VHWPR power cable.
- Case Compatibility: Lists compatible cases, excluding ASUS PRIME series but including ProArt cards and all SKUs with the same dimensions, such as RTX 4080 SUPER. Palit’s upcoming Innity 3 series and ZOTAC’s 4070 BLOWER design are also included.
- Exclusion of RTX 4060: Not included as it is considered a mainstream solution, not powerful enough for the Enthusiast program.
This is NVIDIA’s first attempt to standardize GPU design, aiming for better adherence to specifications and potentially more standardized, smaller GPU designs in the future.
Availability
NVIDIA ACE NIM microservices, including Riva and Audio2Face, are now in production for server deployments, with NVIDIA AI Enterprise software providing enterprise-class support.
Developers can register for early access to ACE NIM microservices on RTX AI PCs.
At the event, Jensen Huang, Founder and CEO of NVIDIA, stated:
Digital humans are set to revolutionize industries. With breakthroughs in multi-modal large language models and neural graphics provided by NVIDIA ACE to our developer ecosystem, we are moving towards a future of intent-driven computing, where interacting with computers feels as natural as interacting with humans.