NVIDIA/nvidia-gcp-samples

NVIDIA/nvidia-gcp-samples

Description: NVIDIA GPU Accelerated Application Samples in Google Cloud Platform

Language: Jupyter Notebook

License: Apache-2.0

Stars: 24

Forks: 12

Open issues: 1

Created: 2020-10-21T21:36:53Z

Pushed: 2026-06-09T00:31:42Z

Default branch: master

Fork: no

Archived: no

README:

NVIDIA GPU Accelerated Application Samples in Google Cloud Platform

Table Of Contents

• [Description](#description)
• [Prerequisites](#prerequisites)
• [Samples](#samples)
• [Additional Resources](#additional-resources)
• [Known issues](#known-issues)
• [License](#license)
• [Maintainers](#maintainers)

Description

This repository maintains sample applications designed for NVIDIA software tools integrated with Google Cloud Platform (GCP), e.g. AI platform, Dataproc, GKE, etc.

For select demonstrations, the sample code will be contained within this repository. For others, we will reference and link to exceptional demonstrations available outside of this repository.

Prerequisites

• Install Google Cloud SDK on your laptop/client workstation, so that gcloud SDK cli interface could be run on the client
• In addition, user could leverage Google Cloud shell

Samples

Deep Learning Inference:

• [Triton autoscaling example with TensorRT optimization in Google Kubernetes Engine](kubernetes-engine-samples/triton_gke)
• [BERT fine tuning, TensorRT optimization, Serve TensorRT engine through Triton in AI Platform Prediction ](ai-platform-samples/bert_on_caip)
• Triton Inference Server application in Google Kubernetes Engine
• Triton GKE Marketplace application, Blog
• AlphaFold batch inference with Vertex AI Pipelines
• Triton in Vertex AI Prediction

Machine Learning and Data Science:

• [XGBoost with LocalCUDACluster Dask single node sample](ai-platform-samples/xgboost_single_node/gcsfs_localcuda)
• RAPIDS XGBoost hyperparameter optimization example
• XGBoost ensemble inference with Triton

Big Data Analytics:

• RAPIDS/Spark on GCP Dataproc
• Churn example with Spark RAPIDS on GCP Dataproc
• TensorRT intergration with Dataflow
• [TensorRT Bert Q&A inference in GCP Dataflow](dataflow-samples/bert-qa-trt-dataflow)
• BigQuery analytics with Dask on GPU

End to End Deep Learning:

• Building a Computer Vision Service Using NVIDIA NGC and Triton in Google Cloud
• NVIDIA Merlin recommender system on GCP Vertex AI
• AutoML Videl Edge on NVIDIA GPU

NIM Microservice and NeMo:

• LLM NIM on Google Kubernetes Engine
• LLM NIM on GCP Vertex AI Workbench
• LLM NIM on GCP Vertex AI Colab Enterprise
• LLM NIM on GCP Cloud Run
• NeMo Retriever NIM on GCP Vertex AI Workbench

Additional Resources

See the following resources to learn more about NVIDIA NGC and GPU resources in Google Cloud Platform

Documentation

• GPU in Google Cloud Platform
• Optimize GPU Performance in Google Cloud Platform
• Getting started with NGC on Google Cloud Platform
• DL Frameworks GPU Performance Optimization Recommendations
• Multi-Instance GPU User Guide for A100 GPU

Known Issues

NA

Contributions

Contributions are welcome. Developers can contribute by opening a pull request and agreeing to the terms in [CONTRIBUTING.MD](CONTRIBUTING.md).

License

See [LICENSE](LICENSE).

Maintainers

• Fortuna Zhang (github: FortunaZhang)
• Dong Meng (github: mengdong)
• Rajan Arora (github: roarjn)
• Ethem Can (github: ethem-kinginthenorth)
• Arun Raman (github: arunraman)