In this presentation, we delve into the cutting-edge realm of large-scale AI training and inference, focusing on the open models and their deployment on Google Cloud Accelerators. Open models such as the Llama family of LLMs and Gemma are state-of-the-art language models that demand robust computational resources and efficient strategies for training and inference at scale. This session aims to provide a comprehensive guide on harnessing the power of PyTorch on Google Cloud Accelerators, specifically designed to meet the high-performance requirements of such models.
Click the blue “Learn more” button above to tap into special offers designed to help you implement what you are learning at Google Cloud Next 25.
talk-data.com
Topic
PyTorch
deep_learning
machine_learning
neural_networks
3
tagged
Activity Trend
16
peak/qtr
2020-Q1
2026-Q1
Top Events
O'Reilly AI & ML Books
7
Databricks DATA + AI Summit 2023
6
O'Reilly Data Science Books
5
PyConDE & PyData Berlin 2023
4
Data Engineering Podcast
4
PyData Paris 2025
4
O'Reilly Data Engineering Books
3
Google Cloud Next '24
3
Computer Vision - Classification d'images avec PyTorch
3
Google Cloud Next '25
3
Introduction à PyTorch
3
PyData Berlin 2025
2
Filtering by:
Google Cloud Next '24
×
by
Nisha Mariam Johnson
(Google Cloud)
,
Sapir Weissbuch
(Lightricks)
,
Philipp Schmid
(Google DeepMind)
,
Milad Mohammadi
(Google Cloud)
More generative AI models are built on PyTorch than on any other framework. We partner with Lightricks to share how PyTorch/XLA offers a performant, automatic compiler experience with all the ease-of-use and ecosystem benefits of PyTorch. Learn from Hugging Face as they share more about the latest features that improve PyTorch/XLA performance and usability on GPUs and TPUs.
Click the blue “Learn more” button above to tap into special offers designed to help you implement what you are learning at Google Cloud Next 25.
by
Paresh Kharya
(Google Cloud)
,
Andrew Clare
(Nuro)
,
Ildoo Kim
(Kakao Brain)
,
Maulin Patel
(Google Cloud)
Deploying AI to production can be bafflingly complex. Learn how Google Cloud is bringing its over two decades of expertise in productionizing planet scale AI to our cloud customers with the AI Hypercomputer architecture. It’s a groundbreaking supercomputing architecture built on performance-optimized hardware (TPUs, GPUs), open software (PyTorch, Jax, Kubernetes), and tailored consumption models that optimize efficiency and productivity across AI training, tuning, and serving. Plus, gain valuable insights from our customers Kakao Brain and Nuro on their journey to deploying large scale AI on Google Cloud.
Click the blue “Learn more” button above to tap into special offers designed to help you implement what you are learning at Google Cloud Next 25.