NVIDIA’s GeForce RTX AI PC Tour in Bangalore, an event to showcase their latest hardware and software technologies, was a testament to how far we’ve come in integrating artificial intelligence into our daily computing experiences. From gaming and content creation to professional applications, NVIDIA pulled back the curtain on its latest advancements in AI hardware and software technologies. NVIDIA didn’t announce anything new in particular but this was more of an opportunity to help experience the things that are usually only seen at major tech expos.
The event was centred around NVIDIA’s GeForce RTX AI PCs, machines that harness the power of Tensor Cores—dedicated AI accelerators present in all GeForce RTX GPUs. These PCs are designed to meet the growing demand for advanced AI capabilities across various domains. With over 100 million GeForce RTX and NVIDIA RTX GPUs already in users’ hands worldwide, NVIDIA claims that they are pushing the envelope to deliver AI acceleration in more than 600 games and applications.
One of the standout demonstrations was that of ChatRTX. This demo app allows users to personalise a GPT large language model (LLM) by connecting it to their own content—be it documents, notes, images, or other data. Leveraging retrieval-augmented generation (RAG), TensorRT-LLM, and RTX acceleration, ChatRTX enables you to query a custom chatbot and receive contextually relevant answers in real-time. We had the opportunity to mess around with ChatRTX firsthand. ChatRTX could easily pull out individual references from some of the technical documents that were present on the machine and even explain some of the concepts within those documents without having to connect to the internet. ChatRTX relies on the prowess of the selected AI model so if an LLM is selected, it can easily comb through text but images will be a problem. But, you can always switch to an image model such as CLIP and then ChatRTX can sift through all your images and locate certain photos based on your prompts. The responses were not swift but they were fast enough, much faster than combing through hundreds of PDFs of images. The fact that this all runs locally on a Windows RTX PC or workstation adds a layer of security and speed that’s hard to overlook.
Another highlight was Audio2Face, a set of AI-based technologies that generate facial animation and lip-sync purely from an audio source. The demo showcased how simple brushstrokes or audio inputs could be transformed into realistic facial animations, complete with nuanced expressions and lip movements that synced perfectly with the speech. The potential applications for this technology are vast. For game developers and animators, Audio2Face could significantly reduce the time and resources needed to animate characters. For content creators, it opens up new avenues for dynamic storytelling and engagement. We saw a preset model narrate stories, rap and even lament its fate. You get an array of sliders to tweak emotions such as Joy, Anger, Cheekiness, etc. and since the model detects emotions from the text, you can also ramp up the Emotion Detection Range if you feel that the virtual actor isn’t putting its best foot forward.
There can’t be an NVIDIA event without an RTX demo and we saw NVIDIA DLSS 3.5 in action with Ray Reconstruction within Star Wars Outlaws built using Massive Entertainment’s Snowdrop Engine. The game features stunning ray-traced visuals and includes NVIDIA DLSS 3.5 with Ray Reconstruction, ensuring that the graphics are not only realistic but also rendered at high frame rates.
Another fascinating demo was from Perfect World Games’ “Legends.” This NVIDIA ACE and digital human technology demo introduced Yun Ni, a character that can see gamers and identify people and objects in the real world using the computer’s camera powered by ChatGPT-4. What really stood out was that the character could perform world building on the spot and generate in-game lore around the world and the many characters that it had to interact with during the course of the game.
For content creators, the demonstration of ComfyUI was an interesting one. As one of the most popular Stable Diffusion applications, ComfyUI is renowned among advanced users for its flexibility in various workflows. The demo showcased the ability to take a selfie and transform it into multiple superhero representations within seconds, thanks to RTX acceleration for Stable Diffusion. This kind of speed and flexibility can significantly enhance workflows in creative industries, allowing artists and designers to iterate quickly and bring their visions to life with unprecedented efficiency.
In conversations with NVIDIA representatives, it was clear that the company sees immense potential in the Indian market. The focus is on democratising access to advanced AI technologies, ensuring that students, professionals, and enthusiasts alike can harness these tools to innovate and excel.
The NVIDIA GeForce RTX AI PC Tour in Bangalore was more than just a showcase of cutting-edge technology; it was a glimpse into the future of computing. The seamless integration of AI into various aspects of PC usage—from gaming and content creation to professional applications—signals a shift towards more intelligent, responsive, and personalised computing experiences.
