Time: 11:00am – 12:00pm PT
Period: 1 hour
Excessive-fidelity simulations in science and engineering are extensively utilized in industrial, seismic, climate/local weather, and life sciences functions. Nevertheless, conventional simulations stay computationally costly and impractical for actual time functions. They’re discretization dependent, which means they don’t simply assimilate both measured or artificial knowledge from varied sources. On account of fast developments in AI for science and engineering issues, machine studying has assumed an essential complementary position in addressing the vital gaps within the conventional strategies.
NVIDIA Modulus is a physics-based machine studying platform that has a number of state-of-the-art community architectures and knowledge, in addition to PDE pushed AI strategies to unravel actual world science and engineering issues. Numerous efficiency options for each single and multi-GPU/node methods, plus connectivity with a number of NVIDIA toolkits and applied sciences can be found in Modulus. Examples and documentation are offered to make sure seamless studying for college students whereas the researchers can customise the framework by varied APIs.
This webinar will introduce you to functions of machine studying, varied domains of science and engineering, in addition to a deep dive into the code implementation, coaching, answer, and visualization features of physics-ML workflow.
By attending this webinar, you’ll find out about:
- The machine studying functions in science & engineering with physics-ML framework, NVIDIA Modulus.
- How one can prolong/modify Modulus to implement your personal work.
- The structure and performance of Modulus, and efficiency enhancements for knowledge & physics pushed methods.
- How the Modulus framework integrates with different Nvidia toolkits and applied sciences: PySDF (for geometry), DALI™ (for knowledge loading), Triton™ (for inference), Omniverse™ platform (for visualization).
Be a part of us after the presentation for a stay Q&A session with Jianjun Xu, Ph.D., Sr. Options Architect, Amazon Net Providers.