Madhya Pradesh has secured investment proposals worth INR 3,200 crore in a single day during an interactive session held in Bengaluru on August 7-8, 2024. The event, organised by Invest Madhya Pradesh, drew over 500 participants, including leading industrialists and investors from the IT, textiles, aerospace, and pharmaceutical sectors.

Among the significant proposals, Google Cloud announced plans to establish a startup hub and Center of Excellence in the state. Chief Minister Mohan Yadav shared that these initiatives aim to enhance the local skilled workforce. 

Meanwhile, chip giant NVIDIA suggested creating a blueprint to transform Madhya Pradesh into the “Intelligence Capital of India.”

The session also saw participation from more than 15 international diplomatic missions and major companies like Infosys, Cognizant, and TCS, further solidifying Madhya Pradesh’s reputation as a prime investment destination.

Chief Minister Yadav highlighted that the proposals could generate approximately 7,000 new jobs in the state, providing a substantial boost to the local economy.

“Discussions were held with these companies regarding the development of information technology in Madhya Pradesh and their future plans, and I am hopeful that with the kind of positive response that we have got, we will witness many IT companies setting up their campuses in Madhya Pradesh,” he said.

In February, Google had signed an MoU with the Maharashtra government to advance scalable AI solutions in sectors like agriculture and healthcare. The recent proposals in Madhya Pradesh signal a continued commitment from tech giants to invest in India’s technological growth.

The post Google, NVIDIA, and Microsoft to Invest INR 3,200 Crore in Madhya Pradesh appeared first on AIM.

NVIDIA has announced its AI Summit 2024, scheduled between October 23 and 25 at the Jio World Convention Centre in Mumbai, India. The summit will feature three days of presentations, hands-on workshops, and networking opportunities aimed at connecting industry experts and exploring advancements in artificial intelligence.

“With accelerated computing infrastructure, research and AI skilling at scale, India has the potential to become the intelligence capital of the world. The upcoming NVIDIA AI Summit is the first-of-its-kind event with a significant focus on India,” said Vishal Dhupar, managing director (Asia-South), NVIDIA. 

“It promises to be topical, relevant for India and is a must-attend for developers, startups and enterprises,” he added.

1. Blackwell is Coming to India 

NVIDIA chief Jensen Huang is set to attend the event and participate in a fireside chat. There’s a strong possibility that Huang will announce the availability of NVIDIA’s latest Blackwell GPUs for the Indian market. 

According to a report, NVIDIA’s newly announced Blackwell GPUs are expected to begin shipping as early as October. Mumbai-based data center and managed cloud infrastructure provider Yotta Data Services is set to benefit from this early release.

“We are in early talks with Nvidia to source Blackwell GPUs as part of our order, and are in the process of finalising all details,” said Sunil Gupta, the co-founder and chief executive of Yotta.

“We’re looking at procuring around 1,000 Blackwell B200 GPUs by October, which would be equivalent to around 4,000 ‘H100’ GPUs. While a timeline isn’t clear yet, we’re expecting the delivery of Blackwell GPUs before the end of this year, and complete our full existing order by the next fiscal,” Gupta said.

The entire order between Yotta and Nvidia, Gupta said, is worth ~$1 billion.

Notably, the production of Blackwell chips has been delayed by three months or more due to design flaws, which could impact customers such as Meta platforms, Google, and Microsoft, who have collectively ordered tens of billions of dollars worth of the chips.

Earlier this year, Yotta, an elite partner of NVIDIA, received the first shipment of 4,000 GPUs. Yotta plans to scale up its GPU inventory to 32,768 units by the end of 2025. Last year, the company announced that it would import 24,000 GPUs, including NVIDIA H100s and L40S, in a phased manner.

2. Data Centres Loading 

During his last visit to India, Huang announced NVIDIA’s partnership with Reliance to develop a foundational large language model tailored to India’s diverse languages and generative AI needs. However, there have been no updates since. It is likely that Reliance will announce something significant during the summit.

Reliance is working with NVIDIA to build AI infrastructure that is over an order of magnitude more powerful than the fastest supercomputer in India today. NVIDIA will provide access to its most advanced GH200 Grace Hopper Superchip and NVIDIA DGX Cloud, an AI supercomputing service in the cloud. 

Reliance said that it will create AI applications and services for their 450 million Jio customers and provide energy-efficient AI infrastructure to scientists, developers and startups across India.

Similarly, NVIDIA partnered with Tata to build an AI supercomputer powered by the next-generation NVIDIA® GH200 Grace Hopper Superchip. TCS will utilize this AI infrastructure to develop and process generative AI applications. Additionally, TCS announced plans to upskill its 600,000-strong workforce through this partnership. 

In a recent earnings call, TCS reported having over $1.5 billion in its AI pipeline, encompassing 270 projects. 

Last year, Infosys expanded its alliance with NVIDIA to train 50,000 employees on NVIDIA’s AI technology, integrating these tools with Infosys Topaz to create generative AI solutions for enterprises.

Similarly, Netweb Technologies also partnered with NVIDIA to manufacture NVIDIA’s Grace CPU Superchip and GH200 Grace Hopper Superchip MGX server designs. This partnership supports the Make in India initiative by building a local ecosystem to address demands around AI and accelerated computing applications for both government and private enterprises.

3. Partnership with Indian AI Startups 

It’s highly likely that Indian AI startups will also make their presence felt at the event. Earlier this year, Dhupar said that he found Krutrim, Sarvam AI, and Immersio to be the three ‘most-exciting’ AI startups from India.

NVIDIA chief Huang believes the future of AI lies in physical AI. Recently, Bengaluru-based startup Control One launched India’s first physical AI agent for the global market. They released a video showcasing this agent, which responds to voice commands via a unique operating system. Control One has already integrated this OS into a forklift.

Control One is also an NVIDIA Inception Partner, which grants the startup access to cutting-edge GPU technology and crucial expertise for developing and scaling its AI systems.

Recently, another Indian AI startup, KissanAI, got accepted into the NVIDIA Inception Program.

Who knows, NVIDIA might partner with People+ AI’s Open Cloud Compute as well. OCC seeks to create an open network of compute resources, making it easier for businesses, especially startups, to access the compute power they need without being locked into specific cloud providers.

Meanwhile, NVIDIA recently partnered with Thapar Institute of Engineering and Technology to advance AI education and research. 

Through this technical collaboration, the university will offer a formidable infrastructure. Its current 227 petaflops of AI performance will be expanded to over 500 petaflops, tailored for the most demanding AI and deep learning tasks. It might follow suit by partnering with more universities. 

4. Made In India PCs 

Recently, NVIDIA announced a collaboration with six Indian PC builders—The MVP, Ant PC, Vishal Peripherals, Hitech Computer Genesys Labs, XRig, and EliteHubs—to launch Made-in-India PCs equipped with RTX AI. This initiative aims to bring advanced AI technology to Indian gamers, creators, and developers.

“Our vision is deeply rooted in the commitment to India’s future in AI and computing. With India’s AI market projected to reach $6 billion by 2027, the opportunity is immense,” said Dhupar. 

“There is an opportunity for India to be the capital of intelligence. The country has the skill sets and talent that understand how to work with a computer,” he added.

The new PCs are part of NVIDIA’s ongoing commitment to gaming and technological advancement. The inclusion of RTX AI technology in these systems offers gamers enhanced performance and visual experiences. 

5. Partnership with Indian Govt

Given Huang’s previous engagements with Indian leaders, including a meeting with Prime Minister Narendra Modi, his address may also touch upon NVIDIA’s plans for collaboration with India in AI and chip manufacturing. 

The Indian government is actively supporting AI development through the IndiaAI mission, launched in March 2024, which aims to position India as a global AI leader by investing in infrastructure and supporting startups. The mission includes an INR 10,300 crore investment to expand AI infrastructure and make GPUs more accessible. As many as 10,000 GPUs will be made available to startups and a marketplace will be created to benefit R&D facilities and startups.

The post NVIDIA AI Summit India 2024 – 5 Key Things to Expect appeared first on AIM.

NVIDIA has announced its AI Summit 2024, set to take place from October 23 to 25 at the Jio World Convention Centre in Mumbai, India. The summit will feature three days of presentations, hands-on workshops, and networking opportunities, aimed at connecting industry experts and exploring advancements in artificial intelligence.

“It’s going to be an exhilarating experience for the tech community and I’m especially proud to be hosting the summit in India,” said Vishal Dhupar, managing director, Asia South NVIDIA. 

NVIDIA CEO Jensen Huang will make a special visit to India for the event, participating in a fireside chat. Huang’s visit follows last year’s announcement of NVIDIA’s partnerships with Reliance, Tata, and Infosys to support India’s AI startup ecosystem and reskill the IT workforce.

NVIDIA has begun delivering its latest chips, such as the GH200 AI, to Indian partners like Tata Communications and Jio Platforms, which are building AI-cloud infrastructure Tata Communications’ Managing Director and CEO, A.S. Lakshminarayanan, confirmed that the installation process for NVIDIA’s AI Cloud is underway, with a full launch expected by the third quarter of this fiscal year.

The summit will showcase NVIDIA’s latest innovations, including the much-anticipated B100 chip, promising significant improvements over its predecessor, the H100. Attendees can explore NVIDIA’s suite of AI solutions designed to accelerate innovation in fields like healthcare, robotics, and industrial digitalization.

Last year, Infosys expanded its alliance with NVIDIA to train 50,000 employees on NVIDIA’s AI technology, integrating these tools with Infosys Topaz to create generative AI solutions for enterprises.

During his visit to India last year, Huang said “India has lots of data,” touching upon the diversity of languages and dialects. He said, “There’s no reason for India to export data to western companies”. 

He believes India has the capability to make in-house LLMs and foundational models. “You have all of your own data. You have the great talent of computer scientists. You produce more computer scientists than any country on the planet; and have the right infrastructure for producing computer scientists. You have an infrastructure for that, right? It’s called AI – Actual Intelligence,” said Jensen, speaking of IITs. 

However, he said India does lack infrastructure – “not the roads and bridges kind”, but AI infrastructure. He said with NVIDIA supercomputers coming in, that has also been taken care of. “You have everything you need to build and use the AI here. But you need to have infrastructure. Just like electric power plants and steam engines, this is now the production of intelligence,” he stressed. 

Following that the Indian government is actively supporting AI development through the IndiaAI mission, launched in March 2024, which aims to position India as a global AI leader by investing in infrastructure and supporting startups. The mission includes a INR 10,300 crore investment to expand AI infrastructure and make GPUs more accessible.

In March 2024, Yotta received the first shipment of 4,000 NVIDIA H100 GPUs. Yotta plans to scale up its GPU inventory to 32,768 units by the end of 2025, following last year’s announcement of importing 24,000 GPUs, including NVIDIA H100s and L40S, in phases.

The post NVIDIA Chief Jensen Huang to Visit India in October appeared first on AIM.

In a move to scale their global marketing campaigns, the Coca-Cola Company has partnered with WPP and NVIDIA to integrate generative AI capabilities using NVIDIA Omniverse and NVIDIA NIM microservices. 

This collaboration, working through WPP Open X, will leverage NVIDIA‘s technology to personalise and customise brand imagery across over 100 markets worldwide. 

This initiative is part of Coca-Cola’s digital transformation strategy, led by Samir Bhutada, global VP of StudioX Digital Transformation.

WPP announced at SIGGRAPH that Coca-Cola will be among the first to adopt NVIDIA NIM microservices for OpenUSD in its Prod X studio. These include USD Search and USD Code, allowing the creation and manipulation of 3D advertising assets with culturally relevant elements using AI-generated images and prompt engineering.

“Our research with NVIDIA Omniverse for the past few years, and the research and development associated with having built our own core USD pipeline and decades of experience in 3D workflows, enabled us to create a tailored experience for the Coca-Cola Company” said Priti Mhatre, managing director for strategic consulting and AI at Hogarth. 

The post WPP and NVIDIA Omniverse Help Coca-Cola Scale Brand-Authentic Generative AI Content appeared first on AIM.

Open Source platform Hugging Face is offering developers Inference-as-a-Service that will be powered by NVIDIA’s NIM. The new service provides 5x better token efficiency for AI models and allows immediate access to NIM microservices running on NVIDIA DGX Cloud. 

The new inference-as-a-service was announced at the ongoing SIGGRAPH 2024, a premier conference and exhibition on computer graphics and interactive techniques happening at Denver, Colorado. The new service will facilitate developers to deploy powerful LLMs such as Llama 2, Mistral AI models and many more with optimisation from NVIDIA NIM microservices. Hugging Face Enterprise Hub users can access serverless inference for increased flexibility and minimal infrastructure overhead with NVIDIA NIM.

When accessed as a NIM, large models such as the 70-billion-parameter version of Llama 3, will deliver up to 5x higher throughput when compared with off-the-shelf deployment on NVIDIA H100 Tensor Core GPU-powered systems.

The new inference service supports Train on DGX Cloud, an AI training service that is already available on Hugging Face.

The Omnipresent NVIDIA NIM 

NVIDIA NIM is a set of AI microservices, including NVIDIA AI foundation models and open-source community models, that has been optimised for inference with standard APIs. It improves token processing efficiency and enhances the NVIDIA DGX Cloud infrastructure, accelerating AI applications. This setup provides faster, more robust results. 

The NVIDIA DGX Cloud platform is tailored for generative AI, offering developers reliable, accelerated computing infrastructure for faster production readiness. It supports AI development from prototyping to production without requiring long-term commitments. 

Hugging Face Dominates

The new announcement banks on an existing partnership between both tech companies and is only going to foster the developer community further. Interestingly, Hugging Face recently announced its profitability with a 220-member team. They also released SmolLM, a series of small language models. 

The post NVIDIA and Hugging Face Offers Inference-as-a-Service with 5x Token Efficiency for AI Models appeared first on AIM.

NVIDIA has introduced new NVIDIA NIM microservices and the NVIDIA Metropolis reference workflow, significantly advancing generative physical AI. These developments, announced at SIGGRAPH, include three fVDB NIM microservices supporting NVIDIA’s deep learning framework for 3D worlds and USD Code, USD Search, and USD Validate microservices for working with Universal Scene Description (OpenUSD). 

These tools enable developers to integrate generative AI copilots and agents into USD workflows, expanding the capabilities of 3D worlds.

Physical AI with Visual AI Agents

Physical AI, which uses advanced simulations and learning methods, is transforming sectors like manufacturing and healthcare by enhancing the ability of robots and infrastructure to perceive, reason, and navigate. Interestingly, NVIDIA chief Jensen Huang had termed the next wave of AI as Physical AI.  

NVIDIA offers a range of NIM microservices tailored to specific models and industries, supporting speech and translation, vision and intelligence, and realistic animation. Visual AI agents, powered by vision language models (VLMs), are increasingly deployed in hospitals, factories, and cities.

In Palermo, Italy, NVIDIA NIM-powered agents help manage traffic efficiently. They have deployed visual AI agents using NVIDIA NIM to uncover physical insights that help them better manage roadways.

Companies such as Foxconn and Pegatron also use the same to design and operate virtual factories, improving safety and efficiency.

Bridging the Simulation-to-Reality Gap

NVIDIA’s physical AI software, including VLM NIM microservices, facilitates a “simulation-first” approach, crucial for industrial automation projects. These tools enable the creation of digital twins, simulating real-world conditions for better AI model training. 

Synthetic data from these simulations can replace costly and hard-to-obtain real-world datasets, enhancing model accuracy and performance. NVIDIA’s NIM microservices and Omniverse Replicator are key in building these synthetic data pipelines.

The post NVIDIA Brings Physical AI Through NIM Microservices for Digital Environments appeared first on AIM.

NVIDIA has unveiled significant advancements to Universal Scene Description (OpenUSD), aimed at expanding its adoption across various sectors, including robotics, industrial design, and engineering. 

These developments, announced at the ongoing SIGGRAPH event in Denver, are set to enhance the capabilities of developers in creating highly accurate virtual worlds, crucial for the next evolution of AI technologies.

The new offerings include NVIDIA NIM Microservices, which allow AI models to generate OpenUSD language for various applications such as answering user queries, generating OpenUSD Python code, and understanding 3D space and physics. 

These services are designed to accelerate the development of digital twins and other virtual environments, providing a more efficient and scalable solution for industries.

Digital World with Generative AI

NVIDIA’s latest generative AI models, available as NIM microservices, are the first of their kind for OpenUSD development. These models facilitate the incorporation of AI copilots and agents into USD workflows, broadening the scope of 3D world-building in sectors like manufacturing, automotive, and robotics. The microservices include USD Code NIM, USD Search NIM, and USD Validate NIM, each designed to streamline the creation and validation of 3D content.

NVIDIA has also announced upcoming NIM microservices, such as USD Layout NIM and USD SmartMaterial NIM, which will further enhance the capabilities of developers working with OpenUSD. These tools are expected to play a pivotal role in the development of next-generation AI applications, particularly in the realm of physical AI and robotics.

OpenUSD Reach with New Connectors

In addition to microservices, NVIDIA introduced a series of USD connectors aimed at bringing generative AI to more industries. Collaborating with Siemens, NVIDIA is integrating OpenUSD pipelines with Siemens’ Simcenter portfolio, enabling high-fidelity visualisation of complex simulation data. This partnership aims to facilitate better decision-making and collaboration among stakeholders.

Moreover, NVIDIA released a connector from the Unified Robotics Description Format to OpenUSD, allowing seamless integration of robot data across various applications. To support the growing OpenUSD ecosystem, NVIDIA announced the OpenUSD Exchange SDK, which helps developers create robust data connectors.

These advancements underscore NVIDIA’s commitment to revolutionising the creation and interaction with 3D content, paving the way for broader adoption and innovation across multiple industries.

The post NVIDIA’s GenAI Models for OpenUSD to Advance Robotics and Digital Twins appeared first on AIM.

NVIDIA announced a comprehensive suite of services, AI models, and computing platforms aimed at accelerating the development of humanoid robots worldwide. The offerings include NVIDIA NIM Microservices for robot simulation, the NVIDIA OSMO orchestration service for managing robotics workloads, and an AI-enabled teleoperation workflow for training robots with minimal human demonstration data.

“We’re advancing the entire NVIDIA robotics stack, opening access for worldwide humanoid developers and companies to use the platforms, acceleration libraries and AI models best suited for their needs,” said NVIDIA chief Jensen Huang, who believes that the next wave of AI is Physical AI. 

NIM and OSMO

NVIDIA NIM microservices offer pre-built containers powered by NVIDIA inference software, reducing deployment times from weeks to minutes. Two new microservices, MimicGen and Robocasa, enhance simulation workflows for generative physical AI in NVIDIA Isaac Sim, a robotics simulation application built on the NVIDIA Omniverse platform. 

MimicGen generates synthetic motion data from teleoperated data captured by devices like Apple Vision Pro, while Robocasa creates robot tasks and simulation-ready environments in OpenUSD, a universal 3D development framework.

NVIDIA OSMO, a cloud-native managed service orchestrates complex robotics development workflows, significantly reducing deployment and development cycle times. This service allows users to manage tasks like generating synthetic data, training models, and conducting reinforcement learning and software-in-the-loop testing at scale.

Data Capture for Humanoid Robots

Training humanoid robots requires extensive data, traditionally gathered through time-consuming teleoperation. NVIDIA’s AI and Omniverse-enabled teleoperation workflow, demonstrated at the SIGGRAPH conference, allows developers to generate large amounts of synthetic data from minimal human demonstrations. 

This process involves capturing teleoperated demonstrations using Apple Vision Pro, simulating them in NVIDIA Isaac Sim, and using MimicGen to create synthetic datasets. The Project GR00T humanoid foundation model is then trained using both real and synthetic data, significantly reducing development time and costs.

Fourier, a company focused on general-purpose robots, highlighted the benefits of NVIDIA’s technology. “Developing humanoid robots is extremely complex, requiring an incredible amount of real data, tediously captured from the real world,” said Alex Gu, CEO of Fourier. “NVIDIA’s new simulation and generative AI developer tools will help bootstrap and accelerate our model development workflows.”

Powering Humanoids

Through a new NVIDIA Humanoid Robot Developer Program, developers can gain early access to the new offerings as well as the latest releases of NVIDIA Isaac Sim, which is built on Omniverse. 

The program has already attracted some of the biggest names in robotics including Boston Dynamics, 1X, ByteDance Research, and Neura Robotics, and others. 

“Boston Dynamics and NVIDIA have a long history of close collaboration to push the boundaries of what’s possible in robotics. We’re really excited to see the fruits of this work accelerating the industry at large, and the early-access program is a fantastic way to access best-in-class technology,” said Aaron Saunders, chief technology officer of Boston Dynamics. 

With the rise in humanoid projects and major companies investing big on the same, NVIDIA is placing itself as an indispensable part in powering these projects. 

The post NVIDIA’s New Robotics Suite to Power The Next Humanoid Race appeared first on AIM.

NVIDIA has introduced fVDB, a cutting-edge deep-learning framework designed to create AI-ready virtual representations of the real world. Unveiled at SIGGRAPH, fVDB builds upon the OpenVDB library, a standard for simulating and rendering sparse volumetric data like smoke, clouds, and fire. 

This innovation aims to advance spatial intelligence in generative AI, crucial for autonomous vehicles and robots operating in 3D environments.

Advancing 3D Simulations and Training

NVIDIA chief Jensen Huang had mentioned in an earlier interview that someday, every single car will have autonomous capabilities, and interestingly, the company is coming up with developments to accelerate the AV race. 

fVDB addresses the challenges of converting real-world data into virtual environments. It leverages techniques such as neural radiance fields (NeRFs) and lidar to generate massive, real-time rendered environments that are essential for training AI systems. 

This framework represents a major leap forward from previous models, enhancing industries’ ability to utilize digital twins and high-resolution virtual spaces for various applications, including urban planning and climate science.

The fVDB framework introduces several significant advancements, including support for spatial scales up to 4x larger than previous models, a 3.5x increase in performance speed, and enhanced interoperability that allows seamless integration of extensive real-world datasets into full-sized 3D environments. 

It also features 10x more operators than earlier frameworks, combining previously separate functionalities into a unified system. 

fVDB will soon be available through NVIDIA NIM inference microservices, such as fVDB Mesh Generation NIM for creating 3D environments, fVDB NeRF-XL NIM for generating large-scale NeRFs in OpenUSD, and fVDB Physics Super-Res NIM for high-resolution physics simulations.

NVIDIA’s ongoing efforts with OpenVDB, including the introduction of NanoVDB and NeuralVDB, have already set industry standards for high-speed, real-time simulations and extensive dataset handling. With fVDB, NVIDIA continues to push the boundaries of 3D deep learning, expanding the applications of virtual environments across various domains

The post NVIDIA Unveils fVDB Framework for Spatial Intelligence in Autonomous Vehicles and Robots appeared first on AIM.

Brev Dev, a San Francisco-based AI/ML development platform that was recently acquired by NVIDIA, launched Brev 2.5 that allows one to develop custom containers on Brev. 

“One of our biggest feature requests was people wanting to bring their own container, instead of just picking from one of the containers that we support that’s now live. As part of that, we redesigned the front end experience,” said Nader Khalil, co-founder and CEO of Brev.Dev, current director of dev tech at NVIDIA.  

When provisioning a new GPU, the previous process required selecting the GPU first and then choosing what to run on it. Now, with Brev 2.5, the order is reversed. A user can first choose the container they wish to run, and then select the appropriate compute resources. This redesign also allows the company to suggest suitable compute options.

Brev for ML Models 

The startup offers seamless building, training and deployment of ML models on the cloud. The biggest advantage of this platform is the ability to offer various cloud providers to allow one to optimise GPU as per cost-efficiency. 

The announcement only strengthens Brev Dev’s commitment to the developer community with NVIDIA as the main backer. 

The post NVIDIA- Acquired Startup Releases Brev v2.5 Custom Containers appeared first on AIM.

The wait is finally over. NVIDIA has commenced the delivery of its latest GH200 AI chips to Indian partners Tata Communications and Jio Platforms. This is part of NVIDIA’s broader initiative to enhance India’s AI-cloud infrastructure, following its partnerships with Reliance and Tata Group companies, announced in September last year.

The GH200 AI chips, designed for high-performance computing, and data analytics, are being integrated into the AI-cloud infrastructure of Tata Communications and Jio Platforms.

Tata Communications Managing Director and CEO A.S. Lakshminarayanan confirmed that the installation process is underway, with a full launch of the AI Cloud with NVIDIA expected by the third quarter of this fiscal year.

The Indian government has been proactive in its support for AI development, as evidenced by the IndiaAI mission launched in March 2024. This mission aims to make India a global leader in AI by investing in computing infrastructure, fostering innovation, and supporting startups. 

The government’s commitment is further underscored by significant investments, including a INR 10,300 crore allocation to expand AI infrastructure and make GPUs more accessible.

In March 2024, Yotta being an elite partner received the first shipment of the 4000 H100s, a first shipment of the NVIDIA GPUs to India. Yotta plans to scale up its GPU inventory to 32,768 units by the end of 2025. Last year, the company announced that it would import 24,000 GPUs, including NVIDIA H100s and L40S, in a phased manner.

However, acquiring the highly valued NVIDIA GPUs is no mean feat. NVIDIA sells its GPUs through the NVIDIA Partner Program, which includes Registered, Preferred, and Elite categories. According to NVIDIA’s blog post, Elite partners represent the highest level of partnership and the tag is reserved for those demonstrating exceptional commitment.

Gupta added that India could build five GPT-4 models simultaneously using its existing infrastructure. “I have ordered 16,000 [GPUs], so if there are five customers each wanting to make a GPT-4, I can handle their load simultaneously,” said Gupta.

The post NVIDIA Starts Supplying GH200 AI Chips to Tata Communications, Jio Platforms appeared first on AIM.

Researchers at NVIDIA have developed Llama3-ChatQA-2-70B, a new large language model that rivals GPT-4-Turbo in handling long contexts up to 128,000 tokens and excels in retrieval-augmented generation (RAG) tasks. 

The model, based on Meta’s Llama3, demonstrates competitive performance across various benchmarks, including long-context understanding, medium-length tasks, and short-context evaluations.

Read the full paper here

The Llama3-ChatQA-2-70B model boasts several key highlights, including its ability to process contexts up to 128,000 tokens, matching the capacity of GPT-4-Turbo. It demonstrates superior performance in RAG tasks compared to GPT-4-Turbo and delivers competitive results on long-context benchmarks extending beyond 100,000 tokens. 

Additionally, the model performs strongly on medium-length tasks within 32,000 tokens and maintains effectiveness on short-context tasks within 4,000 tokens.

The researchers employed a two-step approach to extend Llama3-70B’s context window from 8,000 to 128,000 tokens. This involved continued pre-training on a mix of SlimPajama data with upsampled long sequences, followed by a three-stage instruction tuning process.

Evaluation results show that Llama3-ChatQA-2-70B outperforms many existing state-of-the-art models, including GPT-4-Turbo-2024-04-09, on the InfiniteBench long-context tasks. The model achieved an average score of 34.11, compared to GPT-4-Turbo’s 33.16.

For medium-length tasks within 32,000 tokens, Llama3-ChatQA-2-70B scored 47.37, surpassing some competitors but falling short of GPT-4-Turbo’s 51.93. On short-context tasks, the model achieved an average score of 54.81, outperforming GPT-4-Turbo and Qwen2-72B-Instruct.

The study also compared RAG and long-context solutions, finding that RAG outperforms full long-context solutions for tasks beyond 100,000 tokens. This suggests that even state-of-the-art long-context models may struggle to effectively understand and reason over such extensive inputs.

This development represents a significant step forward in open-source language models, bringing them closer to the capabilities of proprietary models like GPT-4. The researchers have provided detailed technical recipes and evaluation benchmarks, contributing to the reproducibility and advancement of long-context language models in the open-source community.

The post NVIDIA’s New Model ChatQA-2 Rivals GPT-4 in Long Context and RAG Tasks appeared first on AIM.

AI Forum, in partnership with AIM & NVIDIA, is set to host an insightful workshop on August 9, focusing on the optimisation of retrieval-augmented generation (RAG) models to achieve enterprise-grade accuracy. 

This virtual workshop, scheduled from 3:00 to 4:30 pm IST, aims to provide participants with advanced techniques and best practices for enhancing the performance of RAG models.

Meet the Expert – Sagar, NVIDIA:

The workshop, titled ‘Optimising RAG Models for Enterprise-Grade Accuracy: Advanced Techniques and Best Practices’, will be conducted by Sagar Desai, a senior solutions architect specialising in LLMs.

Sagar also specialises in production deployment using NVIDIA’s stack. His expertise encompasses multimodal chatbots, RAG models, and LLM inferencing, ensuring scalable, reliable, and secure AI solutions. 

Sagar’s proficiency in model fine-tuning, including techniques like SFT, PEFT, and RLHF, along with his experience in designing scalable architectures using containerization, makes him a leading authority in the field. His work focuses on achieving state-of-the-art results in GenAI technologies for enterprise-level adoption.

REGISTER NOW

What You Will Learn?

Everyone is talking about how to RAG in the era of generative AI and LLMs. Learning how to do it is definitely the need of the hour for every data and AI professional.

Sagar will leverage his extensive expertise to guide attendees on how to unlock the full potential of RAG models through methods such as query writing, embedding fine-tuning, and reranking strategies. 

Participants can expect to gain valuable insights into scaling RAG models for high accuracy and reliability in real-world applications. The session will also cover best practices for deploying RAG models in production environments, providing a comprehensive understanding of the intricacies involved in optimising these models.

Key Takeaways

• Optimisation of RAG models for enterprise-grade accuracy using advanced techniques.
• Scaling RAG models for high accuracy and reliability in real-world applications.
• Best practices for deploying RAG models in production environments.

Why Attend?

The workshop is designed for technical professionals with a background in natural language processing, machine learning, or AI. 

While a basic understanding of RAG models and LLMs is recommended, prior experience with query writing, embedding fine-tuning, and reranking strategies is not required. 

In addition, having an account on NVIDIA’s Build portal will be beneficial for API calls to models during the workshop. The NVIDIA Developer Program supports developers with essential resources to drive technological innovation.

• Advanced Tools & Technology: Access over 150 SDKs, including the CUDA Toolkit and NVIDIA NIM.
• Community Support: Peer and expert assistance for collaborative problem-solving.
• Hardware Grants: Available for qualified educators and researchers.
• Training Resources: Comprehensive materials to enhance skills.
• Free Software: GPU-optimised tools for AI, HPC, robotics, and more.
• Academic Support: Teaching Kits, Research Grants, and Fellowships.
• Startup Accelerator: NVIDIA Inception offers training, hardware discounts, and networking for AI and data science startups.

Mandatory Pre-requisites for the Workshop

• This session is designed for technical professionals with a background in natural language processing, machine learning, or artificial intelligence. Attendees should have a basic understanding of RAG models and Large Language Models (LLMs). Prior experience with query writing, embedding fine-tuning, and reranking strategies is not required.
• Having an account on the – NVIDIA Developer Forum will help, it will be used for API call to models.
• Python 3.10, Jupyter notebook setup

You can join us for Live Q&A here.

REGISTER NOW

Secure your spot for the AIM Workshop on August 9, 3:00 to 4:30 pm. Don’t miss this opportunity to enhance your understanding of RAG models and their enterprise applications with insights from an industry expert.

REGISTER NOW

The post AI Forum, AIM & NVIDIA Present: Masterclass on Optimizing RAG Models for Enterprise-Grade Accuracy appeared first on AIM.

San Jose-headquartered company Arrcus Inc. successfully raised $30 million in funding from a consortium of investors led by Nvidia Corp., aimed at bolstering its platform designed to streamline data traffic management for enterprises.

Arrcus boasts a customer base that includes SoftBank Corp. and Target Corp., and specialises in optimising data flow across networks crucial for enterprises, data centres, and users. The startup disclosed plans to utilise the capital infusion to scale its global operations and further enhance its technology infrastructure.

Arrcus’ ACE platform, powered by NVIDIA BlueField DPU, employs a distributed microservices architecture that offers flexibility, high performance, scalability, full programmability, modularity, and readiness for hybrid cloud environments.

It supports diverse deployment options such as data processing units (DPUs), merchant silicon, and compute infrastructure. This enables a wide array of applications including low-latency data centre networking, modern edge setups, telecommunications access and transport, and hybrid multi-cloud connectivity.

“We are thrilled to welcome NVIDIA as our latest investor and look forward to building on our collaboration. Arrcus’ leading networking software coupled with NVIDIA’s AI infrastructure will help deliver maximum efficiency to customers from data centres as well as their edge and cloud computing environments,” said Arrcus CEO and chairman, Shekar Ayyar. 

India Presence

Founded in 2016, the startup has around 150 employees and a significant number of them work out of the startup’s campus in Bengaluru. A few years ago, the Bengaluru team expanded its support and engineering team. 

Saudi Aramco’s venture arm Prosperity 7 Ventures, Hitachi Ventures, and General Catalyst, were some of the prominent backers of this recent funding round. 

It’s interesting to note that NVIDIA has been investing and acquiring startups that are providing solutions to enterprises through development or networking platforms. This week, NVIDIA acquired San Francisco-based AI development platform Brev. 

The post NVIDIA Invests in Arrcus, Networking Software Startup with Bengaluru Footprint appeared first on AIM.

Sphere, a new entertainment medium in Las Vegas, is powered by 150 NVIDIA RTX A6000 GPUs, driving the 16x16K displays across its interior and the 1.2 million programmable LED pucks on its exterior, the Exosphere, which is the world’s largest LED screen. 

Robust network connectivity is maintained by NVIDIA BlueField DPUs, NVIDIA ConnectX-6 Dx NICs, NVIDIA DOCA Firefly Service, and NVIDIA Rivermax software, ensuring synchronized content delivery across all display panels.

“Sphere is captivating audiences not only in Las Vegas, but also around the world on social media, with immersive LED content delivered at a scale and clarity that has never been done before,” said Alex Luthwaite, senior vice president of show systems technology at Sphere Entertainment. “This would not be possible without the expertise and innovation of companies such as NVIDIA that are critical to helping power our vision, working closely with our team to redefine what is possible with cutting-edge display technology.”

Sphere, named one of TIME’s Best Inventions of 2023, hosts Sphere Experiences, concerts, residencies from world-renowned artists, and premier corporate events. Rock band U2 opened Sphere with a 40-show run that concluded in March. 

The Sphere Experience featuring Darren Aronofsky’s Postcard From Earth showcases the venue’s technologies, including high-resolution visuals, concert-grade sound, haptic seats, and atmospheric effects.

Sphere Studios, located in Burbank, Calif., creates video content for the venue, transferring it digitally to Sphere in Las Vegas. 

Content is streamed in real time to rack-mounted workstations with NVIDIA RTX A6000 GPUs, delivering three layers of 16K resolution at 60 frames per second. NVIDIA Rivermax software accelerates media streaming, eliminating jitter and optimizing latency.

NVIDIA BlueField DPUs ensure precision timing through the DOCA Firefly Service, synchronizing clocks in a network with sub-microsecond accuracy. “The integration of NVIDIA RTX GPUs, BlueField DPUs and Rivermax software creates a powerful trifecta for modern accelerated computing, supporting the unique high-resolution video streams and strict timing requirements needed at Sphere,” said Nir Nitzani, senior product director for networking software at NVIDIA.

Sphere Studios also developed the Big Sky camera system, capturing uncompressed 18K images from a single camera, eliminating the need to stitch multiple camera feeds together. The studio’s custom image processing software runs on Lenovo servers powered by NVIDIA A40 GPUs, which fuel creative work such as 3D video, virtualisation, and ray tracing. The team uses apps like Unreal Engine, Unity, Touch Designer, and Notch for developing visuals for different shows.

The Las Vegas Sphere is establishing itself among legendary circular performance spaces like the Roman Colosseum and Shakespeare’s Globe Theater. The venue features eye-popping LED displays covering nearly 750,000 square feet inside and outside the venue.

The post Las Vegas Sphere’s Visual Wonders, Powered by NVIDIA appeared first on AIM.

NVIDIA has announced a collaboration with six Indian PC builders—The MVP, Ant PC, Vishal Peripherals, Hi Tech Computer Genesys Labs, XRig, and EliteHubs—to launch ‘Made in India’ PCs equipped with RTX AI. This initiative aims to bring advanced AI technology to Indian gamers, creators, and developers.

“Our vision is deeply rooted in the commitment to India’s future in AI and computing. With India’s AI market projected to reach $6 billion by 2027, the opportunity is immense,” said Vishal Dhupar, Managing Director, Asia South at NVIDIA.

“There is an opportunity for India to be the capital of intelligence. The country has the skill sets and talent who understands how to work with a computer,” he added.

The new PCs are part of NVIDIA’s ongoing commitment to gaming and technological advancement. The inclusion of RTX AI technology in these systems offers gamers enhanced performance and visual experiences. 

Content creators and developers also stand to benefit from the high performance and efficiency of RTX studio workstations, making 3D rendering, video editing, and AI-driven content generation more accessible and cost-effective.

When asked on how NVIDIA is ensuring right pricing with improved performance for its GPUs, Dhupar said, “Creators and developers can improve the performance of their tools sometime by upto 100%. This means that by (leveraging our GPUs), they can reduce the cost of computing tools by that much within that time period. The performance will improve by two to four times”.

The initiative highlights the collaborative efforts to drive technological progress and support a growing community of over 120 million creators in India. The partnership is expected to redefine user experiences and contribute significantly to the industry.

The post NVIDIA Partners with Indian PC Builders to Launch ‘Made in India’ PCs with RTX AI appeared first on AIM.

The global data center semiconductor and component market skyrocketed an unprecedented 152 percent in the first quarter of 2024, marking a new milestone, according to a report from Dell’Oro Group.

This explosive growth was fueled by insatiable demand for GPUs and custom accelerators, particularly in the hyperscale cloud sector.

The report revealed that in Q1 2024, NVIDIA led all vendors in component revenues, accounting for nearly half of the reported figures, as supplies of its H100 GPUs improved for both cloud and enterprise markets. Samsung and Intel followed NVIDIA in the rankings.

Looking ahead, strong growth for accelerators is expected to continue into 2024, with GPUs remaining the primary choice for AI training and inference workloads. NVIDIA’s upcoming Blackwell platform is poised to strengthen the firm’s leadership position.

However, the report anticipates that custom accelerators and offerings from other vendors, such as the AMD MI300X/MI325X Instinct and Intel Gaudi3, will gain some market share.

The report also noted that revenues for Smart NICs and DPUs surged more than 50 percent in Q1 2024, driven by strong hyperscale adoption for both AI and non-AI use cases. Storage drives and memory saw significant price increases as vendors aimed to align supply with demand. The three major memory suppliers shifted production capacity from DRAM to AI-focused High Bandwidth Memory (HBM) products.

Baron Fung, Senior Research Director at Dell’Oro Group, highlighted, “Accelerators such as GPUs continue to drive substantial growth, with shipments hitting record highs each quarter. Meanwhile, traditional server and storage component markets returned to positive year-over-year growth as vendors and cloud service providers ramped up purchases in anticipation of robust system demand later this year.”

General-purpose computing components also rebounded strongly following an inventory correction cycle in 2023, experiencing double-digit revenue growth. “Average selling price (ASP) of components has increased significantly from a year ago adding to topline growth,” Fung explained.

“For CPUs, an increasing mix toward fourth- and fifth-generation CPUs, which have more cores and feature sets compared to their predecessors, have commanded higher ASPs.”

As data centers continue to expand and evolve to support the explosive growth of AI and cloud computing, the demand for high-performance semiconductors and components shows no signs of slowing down.

The post NVIDIA Blackwell Solidify Leadership, AMD & Intel to Gain Ground With MI300X & Gaudi3 appeared first on AIM.

NVIDIA chief Jensen Huang is reportedly paranoid about the future of his company. During a recent podcast with Lex Fridman, Perplexity AI chief Aravind Srinivas revealed that he once asked Huang how he handles success and stays motivated. 

Huang replied, “I am paranoid about going out of business. Every day I wake up in a sweat, thinking about how things could go wrong.”

Huang explained that in the hardware industry, planning two years in advance is crucial because fabricating chips takes time. “You need to have the architecture ready,” he said. “A mistake in one generation of architecture could set you back by two years compared to your competitor.”

NVIDIA, once primarily known within the gaming community for its graphics chips, is now the most valuable public company in the world. Shares of the chipmaker climbed 3.6% on Tuesday, raising its market cap to $3.34 trillion, surpassing Microsoft, now valued at $3.32 trillion. 

Earlier this month, NVIDIA hit $3 trillion for the first time, overtaking Apple. NVIDIA shares have risen more than 170% this year and surged further after the company reported first-quarter earnings in May. The stock has increased over ninefold since the end of 2022, coinciding with the emergence of generative AI.

Huang has privately told colleagues that NVIDIA must ensure it doesn’t follow the path of companies like Cisco or Sun Microsystems, which experienced rapid rises and eventual falls. 

Cisco, for example, became the most valuable company by selling routers during the dot-com bubble but hasn’t recovered from the sales drop-off it experienced when its hardware became a widely available commodity. Huang is determined to avoid a similar fate for NVIDIA.

The post Jensen Huang is Paranoid about the Future of NVIDIA appeared first on AIM.

Hewlett Packard Enterprise (HPE) and NVIDIA announced NVIDIA AI Computing by HPE, a portfolio of co-developed AI solutions and joint go-to-market integrations that enable enterprises to accelerate adoption of generative AI.

Among the portfolio’s key offerings is HPE Private Cloud AI, a first-of-its-kind solution that provides the deepest integration to date of NVIDIA AI computing, networking and software with HPE’s AI storage, compute and the HPE GreenLake cloud.

The offering enables enterprises of every size to gain an energy-efficient, fast, and flexible path for sustainably developing and deploying generative AI applications.

Powered by the new OpsRamp AI copilot that helps IT operations improve workload and IT efficiency, HPE Private Cloud AI includes a self-service cloud experience with full lifecycle management and is available in four right-sized configurations to support a broad range of AI workloads and use cases.

All NVIDIA AI Computing by HPE offerings and services will be available through a joint go-to-market strategy that spans sales teams and channel partners, training and a global network of system integrators — including Deloitte, HCLTech, Infosys, TCS and Wipro — that can help enterprises across a variety of industries run complex AI workloads.

Announced during the HPE Discover keynote by HPE President and CEO Antonio Neri, who was joined by NVIDIA founder and CEO Jensen Huang, NVIDIA AI Computing by HPE marks the expansion of a decades-long partnership and reflects the substantial commitment of time and resources from each company.

“Generative AI holds immense potential for enterprise transformation, but the complexities of fragmented AI technology contain too many risks and barriers that hamper large-scale enterprise adoption and can jeopardize a company’s most valuable asset – its proprietary data.

“To unleash the immense potential of generative AI in the enterprise, HPE and NVIDIA co-developed a turnkey private cloud for AI that will enable enterprises to focus their resources on developing new AI use cases that can boost productivity and unlock new revenue streams,” said Neri.

 “Never before have NVIDIA and HPE integrated our technologies so deeply – combining the entire NVIDIA AI computing stack along with HPE’s private cloud technology – to equip enterprise clients and AI professionals with the most advanced computing infrastructure and services to expand the frontier of AI,” Huang added.

The post HPE Announces ‘NVIDIA AI Computing by HPE’ to Accelerate Generative AI Adoption appeared first on AIM.

Apart from NVIDIA becoming the world’s most valuable company, its chief Jensen Huang answers what new applications in AI he is the most excited about going forward. 

Bets Big on Proactive Customer Service 

Jensen believes that the future of customer service is going to change significantly.

“The number one most impactful AI application will probably be customer service,” said Huang, explaining that the important thing about the chatbot and the customer service is the data flywheel that can capture all the conversation and engagement and create more data. 

“Currently, we’re seeing data growing about 10x every five years. I would not be surprised to see data growing 100x every five years because of customer service,” he said, adding that it will help companies collect more data and insights to extract better intelligence and provide better service. 

He further highlighted that it might help to reach a time when companies are able to contact the customer and proactively solve a problem even before it arises. “Just like preemptive maintenance, we’re going to have proactive customer support,” Huang said while earlier mentioning that every company’s business data is its gold mine. 

GenAI is already changing the game for customer service. Today, many companies are leveraging it to supercharge their customer support. 

Recently, Bland AI put up a cool billboard advertising promoting its AI agent that can handle all sorts of phone calls for businesses in any voice, and it created a buzz. 

Automation Anywhere, a leader in AI-driven automation, also launched new AI Agents that can slash the time of process tasks from hours to minutes, increasing business impact up to tenfold in areas like customer service. 

Velocity, a top Indian cash flow-based financing platform launched Vani AI, India’s first AI-based interactive calling solution for financial institutions to help reduce operational costs by 20-30% while enhancing customer experience.

Fractal Analytics, a leading AI solutions provider for Fortune 500 companies, effectively reduced call handling time by up to 15% using its latest innovation, dubbed Knowledge Assist, on AWS.

During a six-month pilot program, nearly 500 knowledge workers in contact centres adopted Knowledge Assist, handling hundreds of thousands of queries monthly and managing complex data from over 10,000 documents across pdf, doc, and ppt formats. The pilot showed a 10-15% reduction in average data retrieval time and a 30% call deflection rate due to self-service capabilities.

Generative AI for Everyone 

NVIDIA’s chief said that GenAI is everywhere and we’re at the beginning of a new industrial revolution. Instead of generating electricity, we’re generating intelligence.

“Recently, using GenAI, we made it possible to make regional weather predictions down to a couple of kilometres. It would have taken a supercomputer about 10,000 times more capability to predict weather down to a kilometre,” he added, saying that he’s also excited about the fact that GenAI is being used to generate chemicals, proteins, and even physics or physical AI.

Huang believes GenAI can help enhance logistics, insurance, and also keep people out of harm’s way. From physical things, biological things, and GenAI for 3D graphics and digital twins, to creating virtual worlds for video games, every industry is involved in GenAI according to him and those that are not, are just not paying attention. 

When asked about his thoughts on how enterprises can make AI that’s more sustainable, Huang said that sustainability has a lot to do with energy and we don’t need to put AI training data centres where the energy grid is already challenged.

“The Earth has a lot more energy, it’s just in the wrong places. We can capture that excess energy, compress it into an AI model, and then bring these AI models back to the society where we could use it,” he said, adding that AI doesn’t care where it went to school.

The Future is Small

Finally, Huang added that while today the computing experience is retrieval-based, in the future, it’s going to be more contextual, more generative, and right there on the device running a small language model. This will dramatically reduce the amount of internet traffic.

“It’ll be much more generative with some retrieval to augment. The balance of computation will be dramatically shifted towards immediate generation. This way of computing is going to save a ton of energy and it’s very sensible,” he said. 

Similarly, Microsoft and Meta also made announcements focused on small language models. 

Huang highlighted that the big idea about the future as working with AIs is prompting, adding, “We’re going to have so many more interesting questions because we’re going to get a lot of answers very quickly.”

When asked how to best help customers and organisations get started today with GenAI, Huang said that users can leverage platforms like Databricks’ Data Intelligence Platform (DIP) and NVIDIA NIMs. 

NIMs (NVIDIA Inference Microservices) are containerized AI microservices designed to accelerate the deployment of GenAI models across various infrastructures.  

It simplifies the creation of GenAI applications such as copilots and chatbots, by providing scalable deployment, advanced language model support, flexible integration, and enterprise-grade security, thereby enabling developers to build powerful AI applications quickly.

“Go get yourself a NIM on DIP,” he said, encouraging people to engage with AI.

“Whatever you do, just start and engage! GenAI is one of those things you can’t learn by watching or reading about. You just learn by doing. It is growing exponentially and you don’t want to wait and observe an exponential trend because in a couple of years you’ll be left so far behind. So, just get on the train, enjoy it and learn along the way!” suggested the NVIDIA chief.

The post What Excites Jensen Huang About the Future of AI?  appeared first on AIM.

NVIDIA has released HelpSteer2, an open-source dataset designed to train state-of-the-art reward models for aligning LLMs with human preferences. The permissively licensed dataset under CC-BY-4.0 contains 10,681 prompt-response pairs annotated across five attributes on a Likert scale by over 1,000 US-based annotators.

Read the full paper here. 

The HelpSteer2 dataset achieves a state-of-the-art 92.0% accuracy on RewardBench’s primary dataset when used to train a reward model with NVIDIA’s 340B Nemotron-4 base model, outperforming all other open and proprietary models as of June 12, 2024. 

It is highly data-efficient, requiring only 10,000 response pairs compared to the millions used in other preference datasets, thus significantly reducing computational costs. 

It enables the training of reward models that can effectively align large language models like Llama 3 70B to match or exceed the performance of models such as Llama 3 70B Instruct and GPT-4 on major alignment metrics. Additionally, it introduces SteerLM 2.0, a novel model alignment approach that leverages multi-attribute reward predictions to train LLMs on complex, multi-requirement instructions.

“High-quality preference data is crucial for aligning AI systems with human values, but existing datasets are often proprietary or of inconsistent quality,” said Zhilin Wang, senior research scientist at NVIDIA. “

HelpSteer2 provides an open, permissively licensed alternative for both commercial and academic use.

The HelpSteer2 dataset is available on the Hugging Face hub, and the code is open-sourced on NVIDIA’s NeMo-Aligner GitHub repository. 

Sentiment Analysis Datasets

HelpSteer2 trains and guides models to behave in ways that people prefer. Additionally, there are many other sentiment analysis models with applications in various fields, helping enterprises accurately understand and learn from their clients or customers. 

Some examples include Amazon product data, the multi-domain sentiment dataset, and Sentiment140.

The post NVIDIA Rolls Out HelpSteer2 Dataset to Align LLMs appeared first on AIM.

At its Data+ AI Summit, Databricks announced an expanded collaboration with NVIDIA to optimise data and AI workloads by integrating NVIDIA CUDA-accelerated computing into the core of Databricks’ Data Intelligence Platform. 

The partnership aims to boost the efficiency, accuracy, and performance of AI development pipelines for modern AI factories, as data preparation, curation, and processing are crucial for leveraging enterprise data in generative AI applications.

Through this broadened alliance, Databricks is adding native support for NVIDIA GPU acceleration on its Data Intelligence Platform. The announcement builds upon the companies’ existing collaboration to enrich enterprises’ experiences across various use cases, from training classical machine learning models to building and deploying generative AI applications and optimising digital twins.

“We’re thrilled to continue growing our partnership with NVIDIA to deliver on the promise of data intelligence for our customers from analytics use cases to AI,” said Ali Ghodsi, Co-founder and CEO at Databricks. “Together with NVIDIA, we’re excited to help every organisation build their own AI factories on their own private data.”

Jensen Huang, founder and CEO of NVIDIA, emphasised the importance of accelerated computing in reducing data processing energy demands for sustainable AI platforms. “By bringing NVIDIA CUDA acceleration to Databricks’ core computing stack, we’re laying the foundation for customers everywhere to use their data to power enterprise generative AI,” Huang stated.

A key aspect of the partnership involves Databricks developing native support for NVIDIA-accelerated computing in its next-generation vectorised query engine, Photon. This integration is expected to deliver improved speed and efficiency for customers’ data warehousing and analytics workloads. Photon powers Databricks SQL, the company’s serverless data warehouse known for its industry-leading price-performance and total cost of ownership (TCO). The collaboration is anticipated to lead to the next frontier of price-performance.

Databricks Shares a Unique Partnership with NVIDIA 

In the backdrop of Databricks’ Data + AI Summit 2024, Anil Bhasin, the vice president of India and SAARC region at Databricks, told AIM that the company’s partnership with NVIDIA—one of its strategic investors–is significant, alongside helping them improve run times using their SOTA GPUs. 

“We’ve always been known as pioneers of the lake house architecture, and now we’ve created a new category called the data intelligence platform. We’ve embedded generative AI in the lake house, which is a unique approach not many companies are taking,” said Bhasin, saying that NVIDIA is aligned with their vision because the future lies in data intelligence platforms.

He said this allows them to serve every use case, ingest data from any source, and maintain unified governance. This strategic differentiation makes their partnership with NVIDIA truly special and aligns perfectly with NVIDIA’s ‘Sovereign AI’ for enterprises. Nobody other than Databricks is enabling this. 

“The ability for us to query in natural language, converting it to SQL on the back end, empowers the business user to gain insights. That is true democratisation,” avered Bhasin, saying their vision is powerful, and not just NVIDIA; many companies believe in Databricks’ long-term vision. 

NVIDIA x Databricks 

Recently, Databricks’ open-source model DBRX became available as an NVIDIA NIM microservice. NVIDIA NIM inference microservices provide fully optimised, pre-built containers for deployment anywhere, significantly increasing enterprise developer productivity by offering a simple, standardised way to add generative AI models to their applications. 

Launched in March 2024, DBRX was built entirely on top of Databricks, leveraging the platform’s tools and techniques, and was trained with NVIDIA DGX Cloud, a scalable end-to-end AI platform for developers.

The Databricks Data Intelligence Platform offers a comprehensive solution for building, evaluating, deploying, securing, and monitoring end-to-end generative AI applications. With Databricks Mosaic AI’s data-centric approach, customers benefit from an open, flexible platform to easily scale generative AI applications on their unique data while ensuring safety, accuracy, and governance.

Today’s announcement follows Databricks’ strategic acquisition of Tabular, a data management startup founded by the original creators of Apache Iceberg and Linux Foundation Delta Lake, the two leading open-source lakehouse formats. By bringing together these key players, Databricks aims to lead the way in data compatibility, ensuring organisations are no longer limited by the format of their data.

Driven by the growing demand for data and AI capabilities, Databricks achieved over $1.6 billion in revenue for its fiscal year ending January 31, 2024, representing more than 50% year-over-year growth.

The expanded partnership between Databricks and NVIDIA underscores the critical role of accelerated computing and optimised data processing in enabling enterprises to harness the power of generative AI effectively and efficiently.

The post Databricks Partners with NVIDIA to Unleash ‘Sovereign AI’ in Enterprise appeared first on AIM.

Shares of major Asian chipmakers and suppliers to Nvidia rallied sharply on Thursday, riding on the coattails of the US chip giant’s surge to a $3 trillion market capitalisation.

Nvidia’s stock jumped on optimism about the booming demand for AI chips, overtaking Apple to become the world’s second-most valuable company after Microsoft.

Taiwan Semiconductor Manufacturing Co (TSMC), the world’s largest contract chipmaker and a key Nvidia supplier, soared nearly 5% to a record high in Taiwan. China’s biggest chipmaker Semiconductor Manufacturing International Corp (SMIC) climbed 4% in Hong Kong, while Japan’s Tokyo Electron, the country’s most valuable chip firm, advanced 4.3%.

“Nvidia’s valuation milestone underscores the explosive growth potential in AI chips as tech giants race to launch more AI products,” said Rajiv Menon, an analyst at CLSA. “This is having a ripple effect across the semiconductor supply chain, lifting shares of Nvidia’s suppliers and rivals.”

Nvidia’s market value has skyrocketed from $1 trillion to $3 trillion in just over a year, fueled by the rising popularity of generative AI tools like ChatGPT. The company makes the most advanced AI processors currently available, putting it in pole position to benefit from the AI boom.

Sentiment was further boosted by upbeat comments from Dutch semiconductor equipment maker ASML Holding, seen as a bellwether for the chip industry. ASML indicated improving demand from its top customers, mainly TSMC.

Other Nvidia suppliers like Foxconn and Japanese chipmakers Advantest, Renesas Electronics and Disco Corp posted gains of 3-5%. Smaller Chinese players Will Semiconductor and NAURA Technology also ticked higher.

“As AI goes mainstream, we expect a multi-year upcycle in chip demand and technology investments,” Menon said. “While Nvidia is leading the charge, the entire semiconductor ecosystem is poised to benefit from this transformative megatrend.”

However, analysts also cautioned that chip stocks have run up significantly and some consolidation is likely in the near term. Geopolitical risks around US-China tensions and supply chain disruptions also remain potential headwinds for the sector.

Indian IT Also Gains

Indian tech stocks rallied after NVIDIA surpassed Apple to hit a $3 trillion market cap, driven by the AI boom. TCS, Infosys and Tata Communications, which have AI partnerships with NVIDIA, were among the top gainers. TCS and Infosys are each training 50,000 employees on AI technologies with NVIDIA’s support. Tata Communications will leverage NVIDIA’s advanced GH200 Grace Hopper chip to build an AI cloud platform in India.

TCS shares rose 1.1% to ₹3,788, while Infosys gained 1.73% to ₹1,454.95. Tata Communications surged nearly 2% to ₹1,783.45. Smaller firms Rashi Peripherals, the main distributor of NVIDIA’s gaming GPUs in India, climbed 2.4%. Netweb Tech, which will manufacture NVIDIA GPU servers locally, hit a 5% upper circuit.

The post Asian Chip Makers Rally to New Heights as NVIDIA Touches 3 Trillion Valuation appeared first on AIM.

The future, where everyone will have a digital twin helping us carry out our day-to-day tasks at work and beyond, is not far.

NVIDIA announced the general availability of NVIDIA ACE generative AI microservices to accelerate the next wave of digital humans. Companies in customer service, gaming, entertainment, and healthcare are at the forefront of adopting ACE, short for Avatar Cloud Engine, to simplify the creation, animation, and operation of lifelike digital humans across various sectors.

During his keynote speech at Computex 2024, NVIDIA chief Jensen Huang introduced NVIDIA Inference Microservices (NIMs), Avatar-based AI agents capable of working in teams to accomplish missions assigned by humans.

Further, he said that NIMS-based agents will be capable of performing various tasks such as retrieving information, conducting research, or using different tools. 

“NIMs could also use tools that run on SAP and require learning a particular language called ABAP. Other NIMs might perform SQL queries. All of these NIMs are experts assembled as a team,” said Huang.

NVIDIA has made available its suite of ACE digital human GenAI tools, including Riva, which has ASR, TTS, and NMT capabilities for speech recognition and translation, Nemotron LLM for language understanding, Audio2Face for facial animation, and Omniverse RTX for realistic skin and hair rendering.

The Good Side of AI Avatars 

Most recently, Zoom chief Eric Yuan said that he wants users to stop having to attend Zoom meetings themselves.

He believes one of the major benefits of AI at work will be the ability to create what he calls a “digital twin” — a deep fake avatar of yourself that can attend Zoom meetings on your behalf and even make decisions for you, freeing up your time for more important tasks like spending time with your family.

Meanwhile, LinkedIn co-founder Reid Hoffman recently created an AI twin of himself and discussed various topics on AI in an interview with it. He said he deepfaked himself to see “if conversing with an AI-generated version of myself can lead to self-reflection, new insights into my thought patterns, and deep truths.”

Similarly, executive educator and coach Marshall Goldsmith is creating an AI-powered virtual avatar of himself as a one-of-a-kind endeavour to share his skills and preserve his legacy for years. MarshallBoT, an AI-powered virtual business coach, is based on GPT-3.5 from OpenAI.

Aww Inc, a virtual human company based in Japan, introduced its inaugural virtual celebrity, Imma, in 2018. Since then, Imma has become an ambassador for prominent global brands in over 50 countries.

Building on this success, Aww is now poised to integrate ACE Audio 2Face microservices for real-time animation, enabling a highly engaging and interactive communication experience with its users.

OpenAI recently launched GPT-4o, featuring a voice function that makes it ideal for voice-controlled computing. In a new demo, the company demonstrated the model’s ability to generate multiple voices for different characters. Additionally, NVIDIA showcased GPT-4o’s capabilities by creating a digital human that interacted seamlessly with a real person.

“We’ve had the idea of voice control computers for a long time. We had Siri, and we had things before that; they’ve never felt natural to me to use,” said OpenAI’s Sam Altman in a recent podcast. He  used the term ‘model fluidity’ to describe GPT-4o’s capabilities, which lets users ask it to sing, talk faster, use different voices, and speak in various languages.  

Meanwhile, Meta and Apple have also developed photorealistic avatars for Quest and Vision Pro, respectively. The potential for AI avatars as NPCs in the gaming industry is immense, allowing players to interact with them using natural language to enhance their experience. It would be like creating a virtual world within the game. 

AI is Becoming More Intelligent Than Humans

Recently, a video went viral demonstrating a reverse Turing test. The experiment took place in a VR train compartment with five passengers – four AI and one human. The passengers included AI representations of Aristotle, Mozart, Leonardo da Vinci, Cleopatra, and Genghis Khan. Their task was to determine which among them was human through a discussion.

As the conversation progressed, Genghis Khan’s responses focused solely on conquest, lacking the expected nuance of a historical figure. The AI passengers quickly identified this discrepancy, their algorithms detecting the superficiality in his answers.The reverse Turing test is becoming increasingly relevant as AI systems become more sophisticated and capable of convincingly mimicking human behaviour. This is where tools like Worldcoin come into the picture creating a secure, privacy-preserving digital identity that does not store personal information, but rather a cryptographic hash of the biometric data.

The post NVIDIA ACE is Making Digital Avatars Scarily Good  appeared first on AIM.

Narendra Sen, the CEO of NeevCloud, is a man of ambition. He envisions constructing an AI cloud infrastructure tailored for Indian clients, comprising 40,000 graphics processing units (GPUs) by 2026. This infrastructure aims to support Indian enterprises with training, inference, and other AI workloads.

Yotta, another prominent AI cloud provider, has recently gained attention for its recognition as the inaugural NVIDIA Partner Network (NPN) cloud partner in India, achieving the Elite Partner status globally.

Yotta’s ambitious plan is to set up an AI cloud infrastructure with 32,768 GPUs by the end of 2025. NeevCloud wants to do better.

Moreover, NeevCloud will soon launch an AI inferencing platform which will provide open-source models like Llama 3 series, Mistral models and DBRX by Databricks. 

“Later this month, we plan to introduce the DBRX because we see a great demand for the model in this country. Subsequently, we will launch the text-to-image models from Stability AI.

“We are focusing on a step-by-step approach to ensure smooth operation. User experience is paramount, and if everything proceeds as planned, we might expand the range. Meanwhile, we are also enhancing our capabilities in tandem with these developments,” Sen told AIM.

AI Inference 

Inferencing involves applying the pre-trained model to new input data to make predictions or decisions. So far, around 3,500 developers have signed up to use NeevCloud’s inference platform.

The company plans to launch the beta version of the platform this month and provide developers with free tokens to lure them to come and test NeevCloud’s inference platform.

For the inferencing platform, the company plans to leverage around 100 NVIDIA GPUs, including NVIDIA H100s, A100, and L40 GPUs. Moreover, NeevCloud plans to introduce AMD’s M1300X to the mix. 

“AMD’s M1300X provides cost benefits compared to the rest. Indians don’t care about the brand. What they care about is that the API should work, latency should be fast, and they should get immediate tokens – that’s it,” Sen stated.

NeevCloud could also be the first company to bring Groq’s language processing units (LPU) to India. Recently, Sen posted a photo of him on LinkedIn, posing with Jonathan Ross in Groq’s headquarters in San Francisco. 

( Source: LinkedIn) 

Sen, however, refrained from revealing much in this regard as things are still not finalised. “While we all know the big names like NVIDIA, there are other players in the market that we are trying to onboard, like SambaNova,” he said.

AI Infrastructure as a Service 

For the AI cloud, Sen revealed that they have placed an order with HP Enterprise (HPE) for 8000 NVIDIA GPUs, which they are expecting to receive in the second half of this year. 

NeevCloud will compete directly with Yotta in the AI-as-a-infrastructure space. Notably, another company that has already making GPUs more accessible in India is E2E Network. 

The NSE-listed company offers NVIDIA’s H100 GPU and NVIDIA A100 Tensor Core GPUs at a competitive price compared to large hyperscalers. 

Recently, Bhavish Aggarwal, the founder of Ola Cabs, announced his decision to offer Krutrim AI Cloud to Indian developers. 

Similarly, Tata Communications also partnered with NVIDIA to build a large-scale AI cloud infrastructure for its customers in the private as well as the public sector.

While Krutrim and Tata Communications have not revealed the number of GPUs they plan to deploy, NeevCloud plans to deploy another 12,000-15,000 GPUs by 2025. “Then, by 2026 we will deploy the remaining to hit the 40,000 GPUs target,” Sen said.

How will NeevCloud Fund the 40,000 GPU Acquisition?

However, deploying a GPU cluster of around 40,000 GPUs will cost billions of dollars. According to Sen, it will cost them approximately $1.5 billion. 

While Yotta is backed by the Hiranandani Group, NeevCloud is banking on its data centre partners to help not only procure but deploy the GPUs as well.

So far, NeevCloud has partnered with three large data centre companies in India, two of which are based in Chennai and one in Mumbai. One among them is one of the largest data centre operators in India, Sen said.

“What we have in place is a revenue-sharing model. While they already have the data centre infrastructure, they need to deploy the GPUs on our behalf, and NeevCloud will bring in the customers, who will access both their AI cloud capacity (to be deployed) and their data centre servers,” Sen said.

Sen established NeevCloud in 2023. However, Sen has been running a data centre company in Indore called Rackbank Datacenters for many years now. 

Located in Crystal IT Park, Indore, the data centre is 35,000 sq ft and has a capacity of 32,000+ servers.

The company has innovated a liquid immersion cooling technology, named Varuna, to effectively cool high-density computing hardware utilised for AI, machine learning, and high-performance computing (HPC) tasks. 

This method entails immersing servers and other IT equipment in a dielectric, non-conductive coolant liquid, facilitating direct heat transfer from the components to the liquid.

The post Another Indian Startup is Entering the AI Cloud Space with 40,000 GPUs appeared first on AIM.

CUDA is a wonderful piece of tech that allows you to squeeze every bit out of your Nvidia GPU. However, it only works with NVIDIA, and it’s not easy to port your existing CUDA code to other platforms.

You look for an alternative to CUDA, obviously. 

What are the alternatives to CUDA?

1. OpenCL: An open standard for parallel programming across CPUs, GPUs, and other processors with some performance overhead compared to CUDA.
2. AMD ROCm: An open-source GPU computing platform developed by AMD that allows the porting of CUDA code to AMD GPUs.
3. SYCL: A higher-level programming model based on C++ for heterogeneous processors enabling code portability across CUDA and OpenCL through Intel’s DPC++ and hipSYCL.
4. Vulkan Compute: It is a compute API of the Vulkan graphics framework, enabling GPU computing on a wide range of GPUs with lower-level control.
5. Intel oneAPI: It is a cross-architecture programming model from Intel, including a DPC++ compiler for SYCL, offering an alternative to CUDA for Intel GPUs.
6. OpenMP: It is an API for parallel programming on CPUs and GPUs. It uses compiler directives, and recent versions support GPU offloading as an alternative to CUDA.

Let’s address each with more depth.

1. OpenCL

OpenCL (Open Computing Language) is an open industry standard maintained by the Khronos Group that lets you utilise parallel programming across various platform architectures. 

OpenCL allows you to write a program once, which it can then run on several different processors from different companies like AMD, Intel, and NVIDIA.

This can be useful if you want to use the hardware you already have or if you want to choose the best processor for a specific task, regardless of which company made it.

2. AMD ROCm

ROCm (Radeon Open Compute) is a platform designed by AMD to run code effectively on AMD GPUs. But the best part is that ROCm is open-source and can be accessed by everyone.

One of the most important parts of ROCm is called Heterogeneous-computing Interface for Portability, or HIP. HIP is quite close to CUDA programming in terms of syntax. This means if you know how to program CUDA then there’s no stiff learning curve if you’re switching over. 

There’s even a tool called HIPIFY that can automatically convert CUDA code into code that works with HIP and AMD GPUs, with just a few minor changes required.

3. SYCL

SYCL (pronounced “sickle”) is a higher-level programming model based on standard C++ for heterogeneous processors. SYCL is built on top of the C++ programming language, enabling code portability across OpenCL devices.

The core idea of SYCL is to provide the performance of OpenCL with the flexibility of C++. Good examples of SYCL include Intel’s DPC++ (Data Parallel C++) based on Clang/LLVM that can target CUDA and OpenCL devices.

4. Vulkan Compute

Vulkan’s low-overhead design and close-to-metal nature can enable performance close to and sometimes even exceeding CUDA in many compute workloads. It provides compute shaders to enable GPU computing.

Since Vulkan Compute is a relatively new technology, its ecosystem is still maturing in terms of libraries, tools and language binding. It has a steeper learning curve, especially when graphics interoperability is also used.

However, new Vulkan Compute-focused frameworks like Kompute are emerging to make Vulkan GPU computing more accessible.

While Vulkan Compute can also interoperate with APIs like OpenCL, CUDA and DirectX 12, there are some very specific features like CUDA’s dynamic parallelism, that are not available with Vulkan.

5. Intel oneAPI

oneAPI is an open, unified programming model developed by Intel that aims to simplify development across diverse computing architectures (CPUs, GPUs, FPGAs, and other accelerators).

oneAPI consists of a core set of tools and libraries, including DPC++ language and libraries for deep learning, machine learning and more.

A key goal of oneAPI is to provide an alternative to proprietary models like NVIDIA’s CUDA. It aims to prevent vendor lock-in and allow code portability across Intel, NVIDIA, AMD and other hardware.

Furthermore, case studies have shown up to an 18x speedup for compute-intensive algorithms using oneAPI tools and Intel hardware.

6. OpenMP

Open Multi-Processing, or OpenMP, is an API that supports multi-platform shared-memory parallel programming in C, C++, and Fortran. It has also been used for parallel computing in CPUs.

Recent versions of OpenMP, starting from version 4.0, have introduced support for GPU offloading. This allows OpenMP to be used for GPU computing as an alternative to CUDA.

OpenMP provides a higher level of abstraction compared to CUDA. It handles many low-level details like data movement and kernel launches automatically, which can make it easier to use for some developers. 

CUDA is a proprietary solution from Nvidia and it is fine-tuned to get the most out of Nvidia hardware. So, finding an exact replacement may not be possible as they’ll always have an advantage over any open-source platform. Sure, if you want to run parallel computation over other GPUs, then the given solution will get the job done in the most efficient way possible. 

The post Top 6 Parallel Computing Alternatives to CUDA appeared first on AIM.

Snowflake announced its collaboration with NVIDIA during the Snowflake Summit 2024. This partnership aims to empower customers and partners to develop bespoke AI data applications within Snowflake, leveraging NVIDIA’s AI technology.

This collaboration sees Snowflake integrating NVIDIA AI Enterprise software, incorporating NeMo Retriever microservices into Snowflake Cortex AI, Snowflake’s managed LLM and vector search service. This integration allows organisations to link custom models to varied business data, delivering precise responses seamlessly. 

Additionally, Snowflake Arctic, an open, enterprise-grade LLM, now supports NVIDIA TensorRT-LLM software, enhancing performance. Arctic is also accessible as an NVIDIA NIM inference microservice, broadening developers’ access to its capabilities.

As enterprises strive to maximise AI’s potential, the need for data-driven customisation grows. The Snowflake-NVIDIA collaboration facilitates rapid development of specific AI solutions, benefiting businesses across various sectors.

“Pairing NVIDIA’s full stack accelerated computing and software with Snowflake’s state-of-the-art AI capabilities in Cortex AI is game-changing,” stated Sridhar Ramaswamy, CEO of Snowflake. “Together, we are unlocking a new era of AI where customers from every industry and every skill level can build custom AI applications on their enterprise data with ease, efficiency, and trust.”

“Data is the essential raw material of the AI industrial revolution,” said Jensen Huang, founder and CEO of NVIDIA. “Together, NVIDIA and Snowflake will help enterprises refine their proprietary business data and transform it into valuable generative AI.”

Notable NVIDIA AI Enterprise software capabilities offered in Cortex AI include:

• NVIDIA NeMo Retriever: Provides accurate and high-performance information retrieval for enterprises.
• NVIDIA Triton Inference Server: Facilitates the deployment, running, and scaling of AI inference for various applications on any platform.
• NVIDIA NIM inference microservices, part of NVIDIA AI Enterprise, can be deployed within Snowflake as a native app using Snowpark Container Services. This setup allows organisations to deploy foundational models directly within Snowflake easily.

Quantiphi, an AI-first digital engineering firm and ‘Elite’ partner with both Snowflake and NVIDIA, exemplifies this innovation. Quantiphi’s Snowflake Native Apps, baioniq and Dociphi, are designed to enhance productivity and document processing within specific industries. These apps, developed using the NVIDIA NeMo framework, will be available on Snowflake Marketplace.

The Snowflake Arctic LLM, launched in April 2024 and trained on NVIDIA H100 Tensor Core GPUs, is now available as an NVIDIA NIM. This makes Arctic accessible in seconds, either via the NVIDIA API catalogue with free credits or as a downloadable NIM, offering flexible deployment options.

Earlier this year, Snowflake and NVIDIA expanded their collaboration to create a unified AI infrastructure and compute platform in the AI Data Cloud. Today’s announcements mark significant advancements in their joint mission to help customers excel in their AI initiatives.

The post Snowflake, NVIDIA Join Forces to Enhance Custom AI Applications appeared first on AIM.

NVIDIA is one of the most influential hardware giants in the world. Apart from its much sought-after GPUs, the company also provides free courses to help you understand more about generative AI, GPU, robotics, chips, and more. 

Most importantly, all of these are available free of cost and can be completed in less than a day. Let’s take a look at them. 

Register for the Free Workshop >

1. Accelerating Data Science Workflows with Zero Code Changes

Efficient data management and analysis are crucial for companies in software, finance, and retail. Traditional CPU-driven workflows are often cumbersome, but GPUs enable faster insights, driving better business decisions. 

In this workshop, one will learn to build and execute end-to-end GPU-accelerated data science workflows for rapid data exploration and production deployment. Using RAPIDS-accelerated libraries, one can apply GPU-accelerated machine learning algorithms, including XGBoost, cuGraph’s single-source shortest path, and cuML’s KNN, DBSCAN, and logistic regression. 

More details on the course can be checked here – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+T-DS-03+V1

2. Generative AI Explained

This self-paced, free online course introduces generative AI fundamentals, which involve creating new content based on different inputs. Through this course, participants will grasp the concepts, applications, challenges, and prospects of generative AI. 

Learning objectives include defining generative AI and its functioning, outlining diverse applications, and discussing the associated challenges and opportunities. All you need to participate is a basic understanding of machine learning and deep learning principles.

To learn the course and know more in detail check it out here – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+S-NP-01+V1

3. Digital Fingerprinting with Morpheus

This one-hour course introduces participants to developing and deploying the NVIDIA digital fingerprinting AI workflow, providing complete data visibility and significantly reducing threat detection time. 

Participants will gain hands-on experience with the NVIDIA Morpheus AI Framework, designed to accelerate GPU-based AI applications for filtering, processing, and classifying large volumes of streaming cybersecurity data. 

Additionally, they will learn about the NVIDIA Triton Inference Server, an open-source tool that facilitates standardised deployment and execution of AI models across various workloads. No prerequisites are needed for this tutorial, although familiarity with defensive cybersecurity concepts and the Linux command line is beneficial.

To learn the course and know more in detail check it out here – https://courses.nvidia.com/courses/course-v1:DLI+T-DS-02+V2/

4. Building A Brain in 10 Minutes

This course delves into neural networks’ foundations, drawing from biological and psychological insights. Its objectives are to elucidate how neural networks employ data for learning and to grasp the mathematical principles underlying a neuron’s functioning. 

While anyone can execute the code provided to observe its operations, a solid grasp of fundamental Python 3 programming concepts—including functions, loops, dictionaries, and arrays—is advised. Additionally, familiarity with computing regression lines is also recommended.

To learn the course and know more in detail check it out here – https://courses.nvidia.com/courses/course-v1:DLI+T-FX-01+V1/

5. An  Introduction to CUDA

This course delves into the fundamentals of writing highly parallel CUDA kernels designed to execute on NVIDIA GPUs. 

One can gain proficiency in several key areas: launching massively parallel CUDA kernels on NVIDIA GPUs, orchestrating parallel thread execution for large dataset processing, effectively managing memory transfers between the CPU and GPU, and utilising profiling techniques to analyse and optimise the performance of CUDA code. 

Here is the link to know more about the course – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+T-AC-01+V1

6. Building A Brain in 10 Minutes

This course delves into neural networks’ foundations, drawing from biological and psychological insights. Its objectives are to elucidate how neural networks employ data for learning and to grasp the mathematical principles underlying a neuron’s functioning. 

While anyone can execute the code provided to observe its operations, a solid grasp of fundamental Python 3 programming concepts—including functions, loops, dictionaries, and arrays—is advised. Additionally, familiarity with computing regression lines is also recommended.

To learn the course and know more in detail check it out here – https://courses.nvidia.com/courses/course-v1:DLI+T-FX-01+V1/

7. Augment your LLM Using RAG

Retrieval Augmented Generation (RAG), devised by Facebook AI Research in 2020, offers a method to enhance a LLM output by incorporating real-time, domain-specific data, eliminating the need for model retraining. RAG integrates an information retrieval module with a response generator, forming an end-to-end architecture. 

Drawing from NVIDIA’s internal practices, this introduction aims to provide a foundational understanding of RAG, including its retrieval mechanism and the essential components within NVIDIA’s AI Foundations framework. By grasping these fundamentals, you can initiate your exploration into LLM and RAG applications.

To learn the course and know more in detail check it out here – https://courses.nvidia.com/courses/course-v1:NVIDIA+S-FX-16+v1/

8. Getting Started with AI on Jetson Nano

The NVIDIA Jetson Nano Developer Kit empowers makers, self-taught developers, and embedded technology enthusiasts worldwide with the capabilities of AI. 

This user-friendly, yet powerful computer facilitates the execution of multiple neural networks simultaneously, enabling various applications such as image classification, object detection, segmentation, and speech processing. 

Throughout the course, participants will utilise Jupyter iPython notebooks on Jetson Nano to construct a deep learning classification project employing computer vision models. 

By the end of the course, individuals will possess the skills to develop their own deep learning classification and regression models leveraging the capabilities of the Jetson Nano.

Here is the link to know more about the course – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+S-RX-02+V2

9. Building Video AI Applications at the Edge on Jetson Nano

This self-paced online course aims to equip learners with skills in AI-based video understanding using the NVIDIA Jetson Nano Developer Kit. Through practical exercises and Python application samples in JupyterLab notebooks, participants will explore intelligent video analytics (IVA) applications leveraging the NVIDIA DeepStream SDK. 

The course covers setting up the Jetson Nano, constructing end-to-end DeepStream pipelines for video analysis, integrating various input and output sources, configuring multiple video streams, and employing alternate inference engines like YOLO. 

Prerequisites include basic Linux command line familiarity and understanding Python 3 programming concepts. The course leverages tools like DeepStream, TensorRT, and requires specific hardware components like the Jetson Nano Developer Kit. Assessment is conducted through multiple-choice questions, and a certificate is provided upon completion. 

For this course, you will require hardware including the NVIDIA Jetson Nano Developer Kit or the 2GB version, along with compatible power supply, microSD card, USB data cable, and a USB webcam. 

To learn the course and know more in detail check it out here – https://courses.nvidia.com/courses/course-v1:DLI+S-IV-02+V2/

10. Build Custom 3D Scene Manipulator Tools on NVIDIA Omniverse

This course offers practical guidance on extending and enhancing 3D tools using the adaptable Omniverse platform. Taught by the Omniverse developer ecosystem team, participants will gain skills to develop advanced tools for creating physically accurate virtual worlds. 

Through self-paced exercises, learners will delve into Python coding to craft custom scene manipulator tools within Omniverse. Key learning objectives include launching Omniverse Code, installing/enabling extensions, navigating the USD stage hierarchy, and creating widget manipulators for scale control. 

The course also covers fixing broken manipulators and building specialised scale manipulators. Required tools include Omniverse Code, Visual Studio Code, and the Python Extension. Minimum hardware requirements comprise a desktop or laptop computer equipped with an Intel i7 Gen 5 or AMD Ryzen processor, along with an NVIDIA RTX Enabled GPU with 16GB of memory. 

To learn the course and know more in detail check it out here – https://courses.nvidia.com/courses/course-v1:DLI+S-OV-06+V1/

11. Getting Started with USD for Collaborative 3D Workflows

In this self-paced course, participants will delve into the creation of scenes using human-readable Universal Scene Description ASCII (USDA) files. 

The programme is divided into two sections: USD Fundamentals, introducing OpenUSD without programming, and Advanced USD, using Python to generate USD files. 

Participants will learn OpenUSD scene structures and gain hands-on experience with OpenUSD Composition Arcs, including overriding asset properties with Sublayers, combining assets with References, and creating diverse asset states using Variants.

To learn more about the details of the course, here is the link – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+S-FX-02+V1

12. Assemble a Simple Robot in Isaac Sim

This course offers a practical tutorial on assembling a basic two-wheel mobile robot using the ‘Assemble a Simple Robot’ guide within the Isaac Sim GPU platform. The tutorial spans around 30 minutes and covers key steps such as connecting a local streaming client to an Omniverse Isaac Sim server, loading a USD mock robot into the simulation environment, and configuring joint drives and properties for the robot’s movement. 

Additionally, participants will learn to add articulations to the robot. By the end of the course, attendees will gain familiarity with the Isaac Sim interface and documentation necessary to initiate their own robot simulation projects. 

The prerequisites for this course include a Windows or Linux computer capable of installing Omniverse Launcher and applications, along with adequate internet bandwidth for client/server streaming. The course is free of charge, with a duration of 30 minutes, focusing on Omniverse technology. 

To learn the course and know more in detail check it out here – https://courses.nvidia.com/courses/course-v1:DLI+T-OV-01+V1/

13. How to Build Open USD Applications for industrial twins

This course introduces the basics of the Omniverse development platform. One will learn how to get started building 3D applications and tools that deliver the functionality needed to support industrial use cases and workflows for aggregating and reviewing large facilities such as factories, warehouses, and more. 

The learning objectives include building an application from a kit template, customising the application via settings, creating and modifying extensions, and expanding extension functionality with new features. 

To learn the course and know more in detail check it out here – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+S-OV-13+V1

14. Disaster Risk Monitoring Using Satellite Imagery

Created in collaboration with the United Nations Satellite Centre, the course focuses on disaster risk monitoring using satellite imagery, teaching participants to create and implement deep learning models for automated flood detection. The skills gained aim to reduce costs, enhance efficiency, and improve the effectiveness of disaster management efforts. 

Participants will learn to execute a machine learning workflow, process large satellite imagery data using hardware-accelerated tools, and apply transfer-learning for building cost-effective deep learning models. 

The course also covers deploying models for near real-time analysis and utilising deep learning-based inference for flood event detection and response. Prerequisites include proficiency in Python 3, a basic understanding of machine learning and deep learning concepts, and an interest in satellite imagery manipulation. 

To learn the course and know more in detail check it out here – https://courses.nvidia.com/courses/course-v1:DLI+S-ES-01+V1/

15. Introduction to AI in the Data Center

In this course, you will learn about AI use cases, machine learning, and deep learning workflows, as well as the architecture and history of GPUs.  With a beginner-friendly approach, the course also covers deployment considerations for AI workloads in data centres, including infrastructure planning and multi-system clusters. 

The course is tailored for IT professionals, system and network administrators, DevOps, and data centre professionals. 

To learn the course and know more in detail check it out here – https://www.coursera.org/learn/introduction-ai-data-center

16. Fundamentals of Working with Open USD

In this course, participants will explore the foundational concepts of Universal Scene Description (OpenUSD), an open framework for detailed 3D environment creation and collaboration. 

Participants will learn to use USD for non-destructive processes, efficient scene assembly with layers, and data separation for optimised 3D workflows across various industries. 

Also, the session will cover Layering and Composition essentials, model hierarchy principles for efficient scene structuring, and Scene Graph Instancing for improved scene performance and organisation.

To know more about the course check it out here – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+S-OV-15+V1

17. Introduction to Physics-informed Machine Learning with Modulus 

High-fidelity simulations in science and engineering are hindered by computational expense and time constraints, limiting their iterative use in design and optimisation. 

NVIDIA Modulus, a physics machine learning platform, tackles these challenges by creating deep learning models that outperform traditional methods by up to 100,000 times, providing fast and accurate simulation results.

One will learn how Modulus integrates with the Omniverse Platform and how to use its API for data-driven and physics-driven problems, addressing challenges from deep learning to multi-physics simulations.

To learn the course and know more in detail check it out here – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+S-OV-04+V1

18. Introduction to DOCA for DPUs

The DOCA Software Framework, in partnership with BlueField DPUs, enables rapid application development, transforming networking, security, and storage performance. 

This self-paced course covers DOCA fundamentals for accelerated data centre computing on DPUs, including visualising the framework paradigm, studying BlueField DPU specs, exploring sample applications, and identifying opportunities for DPU-accelerated computation. 

One gains introductory knowledge to kickstart application development for enhanced data centre services.

To learn the course and know more in detail check it out here – https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+S-NP-01+V1

Additional Inputs Contributed – Gopika Raj

The post 18 Free AI Courses by NVIDIA in 2024 appeared first on AIM.

At NVIDIA COMPUTEX 2024, the GPU giant announced the availability of NVIDIA AI Enterprise-IGX with Holoscan on the IGX platform, improving real-time AI computing capabilities for healthcare, industrial, and scientific applications at the edge. This integration enables faster development and deployment of AI solutions with enterprise-grade software and support. 

NVIDIA AI Enterprise-IGX claims to offer high performance, security, and support for the edge computing software stack, streamlining AI-powered operations and the deployment of AI applications at scale. 

The inclusion of Holoscan, a sensor-processing platform, further enhances the development and deployment of AI and high-performance computing applications, delivering real-time insights. The combination of these technologies cuts the time and costs required to build advanced AI solutions across various industries, meeting unique performance and regulatory requirements.

The NVIDIA IGX platform, now supporting the RTX 6000 Ada GPU and the IGX Orin 500 system-on-module, delivers significant improvements in AI performance and computing power.

“As software-defined functionality continues to transform businesses across industries, enterprises are seeking powerful edge AI solutions that can meet their unique performance and regulatory requirements,” said Deepu Talla, vice president of robotics and edge computing at NVIDIA.

Customer Stories

Leading medical technology companies are rapidly adopting NVIDIA IGX with Holoscan. For example, Medtronic uses the platform for its GI Genius intelligent endoscopy module, which is the first FDA-cleared AI-assisted colonoscopy tool. This technology helps physicians detect polyps that could lead to colorectal cancer. 

Moon Surgical employs IGX with Holoscan to power its Maestro System, a surgical robotics system designed to assist surgeons with precision and control during minimally invasive procedures. The platform’s capabilities have accelerated the development and enhancement of these medical technologies, ultimately improving patient care and safety.

In the industrial sector, ADLINK leverages NVIDIA IGX to build industrial-grade edge AI solutions that enhance factory automation and robotic collaboration. These solutions improve functional safety and high-bandwidth sensor processing, transforming operations like machine movement routing, robotic arm operation, and charging-station monitoring. The platform’s powerful computing capabilities ensure more efficient and seamless human-robot collaboration.

The SETI Institute is another notable adopter, using NVIDIA IGX Orin to power radio astronomy capabilities at the Hat Creek Radio Observatory. This technology enables the processing of multiple terabits per second of radio telescope data, facilitating the detection of weaker and rarer astrophysical phenomena. The advanced capabilities of the IGX platform, combined with Holoscan, provide exceptional computational performance for real-time radar processing and radio astronomy.

Clinical Trials Optimised with NVIDIA’s Meta Llama 3

Available as an NVIDIA NIM inference microservice, Meta Llama 3 supports a wide range of applications, including surgical planning, digital assistants, drug discovery, and clinical trial optimisation. 

At COMPUTEX, NVIDIA announced that hundreds of AI ecosystem partners are integrating NIM into their solutions. Over 40 healthcare and life sciences startups and enterprises are using the Llama 3 NIM to build and run applications that accelerate digital biology, digital surgery, and digital health.

Deloitte, for example, uses the Llama 3 NIM and other microservices for drug discovery and clinical trials, driving efficiency in garnering data-based insights from gene to function. 

Transcripta Bio uses Llama 3 for accelerated intelligent drug discovery, leveraging its AI modeling suite, Conductor AI, to predict the effects of new drugs. In clinical trials, companies like Quantiphi and ConcertAI utilise NVIDIA NIM to develop generative AI solutions for research and patient care, enhancing workforce productivity and improving outcomes. 

Mendel AI uses the Llama 3 NIM for its Hypercube copilot, offering a 36% performance improvement in understanding medical data at scale.

Precision medicine company SimBioSys uses the Llama 3 NIM to analyse breast cancer diagnoses and provide tailored guidance for physicians. Artisight automates documentation and care coordination with ambient voice and vision systems, while AITEM builds healthcare-specific chatbots. Abridge uses the NIM for clinical conversation summarisation, improving the efficiency and accuracy of physician-patient encounters.

Recently, the company announced its Q1 FY25 results, reporting a profit of $14.881 billion, a 600% increase from the same quarter last year. The company’s revenue reached $26.04 billion, exceeding the $24.65 billion estimate, and earnings per share (EPS) were $6.12, significantly higher than the projected $5.59.

The post NVIDIA Boosts Real-Time AI for Healthcare with Enterprise-IGX & Holoscan appeared first on AIM.

“Robots need to be able to deal with uncertainty if they’re going to be useful to us in the future. They need to be able to deal with unexpected situations and that’s sort of the goal of a general purpose or multi-purpose robot, and that’s just hard,” said Robert Playter, CEO of Boston Dynamics, in an interview with Lex Fridman last year. 

Playter couldn’t have been more real in describing the difficulty in robotics. Boston Dynamics, which began developing general purpose robots in the early 2000s, introduced its humanoid Atlas only in 2013. Apart from struggling for investments in robotics, training robots is always a challenge. 

Simulation for Robots

Simulated training is the most commonly adopted technique to equip general purpose robots for the real world. This is where virtual environments are created to develop, test and refine algorithms for robots to mimic real-world conditions. 

“Simulation works very well for certain aspects. They work well in simulation for tasks like walking and doing backflips, where you need to balance your robot. And that is the only way,” said Mankaran Singh, founder of Flow Drive, which makes autonomous vehicle capabilities. 

However, for tasks that can be learned through imitation such as folding shirts, it does not require a simulated environment. 

Simulation is Not the Only Way

CynLr Robotics, a Bengaluru-based deep-tech company that is building robotic arms, believes simulation is not the only way to train its robots. “There are so many layers of perception and fundamental intuition using perception that are still missing. These are capabilities that we should focus on to be able to make them more autonomous,” said Gokul NA, founder of CynLr.  

Meanwhile, NVIDIA’s Isaac Sim that is powered by Omniverse, is a robotic simulation platform that provides a virtual environment for AI-based robots to design, test, and train. 

“We do leverage those [Omniverse] technologies as a tool, but you can’t say a tool is the solution,” said Gokul. The limitations come into the picture when you bring these robots into the real world. 

“When you bring from a simulated assumption to reality, it doesn’t work. It doesn’t work at all, because it has never learned that. It has learned something else independently. Your mistakes are what it has learned, what you have left out,” he said. 

He attributes this gap to machines lacking the cognitive layers that aid in understanding objects and environments that can lead to discrepancies between what is seen and what is understood. 

Imitation learning is another common method for training robots where a user can demonstrate a task. However, it also comes with its limitations. For instance, if a user tries to train a robot to pick a white-coloured mug, the robot will fail to pick mugs of other colours.  

Arm and Humanoid Robots 

Similarly the form factor of general purpose robots also has a huge role to play in training them. For instance, robotics arm manufacturing requires a lot of manipulation, something that most companies overlook. 

Gokul believes that today’s robotics developments, especially robotic arms, are more of ‘record and playback machines’ with sophisticated manipulation, however, they lack in perception. “Most cases where you want to commercially deploy these robots, you don’t need legs. Wheels are more than enough, but you need more capability with the hands,” said Gokul, hinting at the current humanoids that are being developed.

2024 being the year of robotics, many players such as Figure AI, Tesla, UniTree, and Aptronix are focusing on building humanoid robots, while Google DeepMind and other research institutes are training and developing arm-based robots to execute multiple functions. 

AutoRT, SARA-RT and RT-Trajectory are a few robotics research systems Google DeepMind released. Stanford University introduced Mobile ALOHA, a system designed to replicate bimanual mobile manipulation tasks necessitating full body control – cooking being the main task demonstrated. 

NVIDIA: The Robot-Enabler

In addition to Omniverse, GPU giant NVIDIA is aggressively investing in robotics and recently unveiled GR00T, a general-purpose foundation model for humanoid robots. Robots powered by GR00T are engineered to understand natural language and mimic human movements by observing action. 

“Building foundation models for general humanoid robots is one of the most exciting problems to solve in AI today,” said NVIDIA chief Jensen Huang, at GTC 2024. 

NVIDIA is even building a comprehensive AI platform for all the leading humanoid robot companies, including OpenAI-backer 1X Technologies, Agility Robotics, Boston Dynamics, Figure AI, Unitree Robotics and many more. 

Not just NVIDIA, other players are also enabling the robot training ecosystem. OpenAI-backed Physical Intelligence which recently raised $70M in funding, is an emerging startup working towards bringing general-purpose AI into the physical world.

The post Real Struggles of Bringing Robots from Simulation to Reality  appeared first on AIM.