k2tf
aws-virtual-gpu-device-plugin
Our great sponsors
k2tf | aws-virtual-gpu-device-plugin | |
---|---|---|
4 | 3 | |
1,142 | 132 | |
- | - | |
2.7 | 0.0 | |
5 months ago | over 1 year ago | |
Go | Jupyter Notebook | |
Mozilla Public License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
k2tf
-
Tool to convert set of yaml ( kustomize generated ) to terraform ?
This might be what you are looking for: https://github.com/sl1pm4t/k2tf
-
HELM vs KUSTOMIZE
... and if you're an opinionated person, like me, and you value consolidated infrastructure atomicity as a whole along side locks for everything. You'd port cherry-picked helm charts as terraform modules with k2tf, and build every docker container from scratch, with forced layer invalidation to perform security updates for every image, using the docker and kubernetes providers respectively.
-
Share a GPU between pods on AWS EKS
Pro tip: If you want to convert k8s yaml files to .tf, you can use k2tf (repo) that is able to convert the resource types of the yaml top their appropriated counterparts of the k8s provider for terraform. To install it, just:
-
Does anyone use terraform to manage Kubernetes objects as opposed to using plain yamls/helm charts/kustomize?
Almost all is created as manifest/helm in K8S world, too much toil to convert (tool like https://github.com/sl1pm4t/k2tf help but exists corners cases)
aws-virtual-gpu-device-plugin
-
Share a GPU between pods on AWS EKS
This project (available here) uses the k8s device plugin described by this AWS blog post to make GPU-based nodes publish the amount of GPU resource they have available. Instead of the amount of VRAM available or some abstract metric, this plugin advertises the amount of pods/processes that can be connected to the GPU. This is controlled by what is called by NVIDIA as Multi-Process Service (MPS).
-
[D] Serverless solutions for GPU inference (if there's such a thing)
AWS has apparently already started using this type of tech as of this year (see lost below). They mention virtual gpus but this particular solution probably won't help OP unfortunately. https://aws.amazon.com/blogs/opensource/virtual-gpu-device-plugin-for-inference-workload-in-kubernetes/
-
AWS open source news and updates No.41
The post explores GPU device plugin to address how to set fractional number of GPU resource for each pod by implementing the Kubernetes device plugin and Nvidia MPS. This project has been open sourced on GitHub.
What are some alternatives?
terraformer - CLI tool to generate terraform files from existing infrastructure (reverse Terraform). Infrastructure to Code
kserve - Standardized Serverless ML Inference Platform on Kubernetes
hcl - HCL is the HashiCorp configuration language.
aws-eks-share-gpu - How to share the same GPU between pods on AWS EKS
terraform-provider-flux - Terraform provider for bootstrapping Flux
k8s-device-plugin - NVIDIA device plugin for Kubernetes
terraform-provider-kubernetes - Terraform Kubernetes provider
determined - Determined is an open-source machine learning platform that simplifies distributed training, hyperparameter tuning, experiment tracking, and resource management. Works with PyTorch and TensorFlow.
asdf-awscli
booster - Software development framework specialized in building highly scalable microservices with CQRS and Event-Sourcing. It uses the semantics of the code to build a fully working GraphQL API that supports real-time subscriptions.