|
1 | 1 | +++ |
2 | 2 | title = "Getting Started with Kubeflow" |
3 | | -description = "Quickly get running with your ML Workflow" |
| 3 | +description = "Get your machine-learning workflow up and running on Kubeflow" |
4 | 4 | weight = 1 |
5 | 5 | +++ |
6 | 6 |
|
7 | | -There are a variety of ways to install Kubeflow |
| 7 | +There are various ways to install Kubeflow. Choose one of the following options |
| 8 | +to suit your environment (cloud, on premises (on prem), or local): |
8 | 9 |
|
9 | | -* GCP users should follow the [guide for deploying on GCP](/docs/gke/deploy/) |
10 | | -* On prem users or users with an existing Kubernetes cluster should follow the [guide for deploying on kubernetes](/docs/started/getting-started-k8s/) |
11 | | -* Users who want to run Kubernetes locally in a virtual machine can select one of the following options |
| 10 | +* To use Kubeflow on Google Cloud Platform (GCP) and Kubernetes Engine (GKE), |
| 11 | + follow the [GCP deployment guide](/docs/gke/deploy/). |
| 12 | +* To use Kubeflow on Amazon Web Services (AWS), |
| 13 | + follow the [AWS deployment guide](/docs/aws/deploy/). |
| 14 | +* If you have an existing Kubernetes cluster or want to use Kubeflow on prem, |
| 15 | + follow the [guide to deploying Kubeflow on |
| 16 | + Kubernetes](/docs/started/getting-started-k8s/). |
| 17 | +* If you want to run Kubernetes locally in a virtual machine (VM), choose one of |
| 18 | + the following options: |
12 | 19 |
|
13 | 20 | * [MiniKF setup](/docs/started/getting-started-minikf/) |
14 | 21 | * MiniKF is a fast and easy way to get started with Kubeflow. |
15 | 22 | * It installs with just two commands and then you are up for |
16 | | - experimentation, and for running complete Kubeflow Pipelines. |
| 23 | + experimentation, and for running complete Kubeflow Pipelines. |
17 | 24 | * MiniKF runs on all major operating systems (Linux, macOS, Windows). |
18 | 25 |
|
19 | 26 | * [Minikube setup](/docs/started/getting-started-minikube/) |
20 | | - * Minikube leverages virtualization applications like [Virtual |
21 | | - Box](https://www.virtualbox.org/) or [VMware |
22 | | - Fusion](https://www.vmware.com/products/fusion.html) to host the virtual |
23 | | - machine and provides a CLI that can be leveraged outside of the VM. |
24 | | - * Minikube defines a fully baked ISO that contains a minimal operating |
25 | | - system and kubernetes already installed. |
| 27 | + * Minikube uses virtualization applications like |
| 28 | + [VirtualBox](https://www.virtualbox.org/) or [VMware |
| 29 | + Fusion](https://www.vmware.com/products/fusion.html) to host the VM |
| 30 | + and provides a CLI that you can use outside the VM. |
| 31 | + * Minikube defines a fully-baked |
| 32 | + [ISO image](https://en.wikipedia.org/wiki/ISO_image) that contains a |
| 33 | + minimal operating system and Kubernetes already installed. |
26 | 34 | * This option may be useful if you are just starting to learn and already |
27 | | - have one of the virtualization applications already installed. |
| 35 | + have one of the virtualization applications installed. |
28 | 36 |
|
29 | | - * [Microk8s setup](/docs/started/getting-started-multipass/) |
30 | | - * The benefits of using [Microk8s](https://microk8s.io/) include: |
31 | | - - Can be installed on any Linux system as a |
32 | | - [snap](https://snapcraft.io/) |
33 | | - - Strong isolation and update semantics - your single-node cluster |
34 | | - will be updated within a short period after upstream Kubernetes |
35 | | - releases. |
36 | | - - GPU pass through built in - e.g. **microk8s.enable gpu** |
37 | | - * If you are not on a Linux machine, or you want to use Kubeflow in a |
38 | | - confined environment, then use |
39 | | - [Multipass](https://github.com/CanonicalLtd/multipass) to launch a |
40 | | - virtual machine. Benefits include: |
41 | | - - Ubuntu [cloud-images](http://cloud-images.ubuntu.com/) already |
42 | | - integrated. |
43 | | - - Lightweight hypervisor using native operating system mechanisms |
44 | | - (e.g. [Hypervisor |
45 | | - Framework](https://developer.apple.com/documentation/hypervisor) on |
46 | | - MacOS, [Hyper-V on Windows |
47 | | - 10](https://docs.microsoft.com/en-us/virtualization/hyper-v-on-windows/quick-start/enable-hyper-v), |
48 | | - QEMU/KVM for linux) |
49 | | - - Eliminates the need to install a separate virtualization |
50 | | - application. |
51 | | - - You can use cloud-init to customize the VM (as you might in a cloud) |
| 37 | + * [MicroK8s setup](/docs/started/getting-started-multipass/) |
| 38 | + * [MicroK8s](https://microk8s.io/) can provide the following benefits: |
| 39 | + - A small, fast, secure, single node Kubernetes installation that installs on any |
| 40 | + Linux system as a [snap](https://snapcraft.io/microk8s). |
| 41 | + - Strong isolation and update semantics - your cluster |
| 42 | + is updated within a short period after upstream Kubernetes |
| 43 | + releases. |
| 44 | + - Built-in support to enable an installed GPU: |
| 45 | + `microk8s.enable gpu` |
| 46 | + * MicroK8s requires Linux. If you are not on a Linux machine, or you want |
| 47 | + to confine your Kubeflow to a disposable machine, the installation guide |
| 48 | + show you how to use |
| 49 | + [Multipass](https://github.com/CanonicalLtd/multipass) to launch a VM. |
| 50 | + Benefits include: |
| 51 | + - [Ubuntu Cloud Images](http://cloud-images.ubuntu.com/) already |
| 52 | + integrated. |
| 53 | + - Lightweight hypervisor using native operating system mechanisms |
| 54 | + (for example, [Hypervisor |
| 55 | + Framework](https://developer.apple.com/documentation/hypervisor) on |
| 56 | + macOS, [Hyper-V on Windows |
| 57 | + 10](https://docs.microsoft.com/en-us/virtualization/hyper-v-on-windows/quick-start/enable-hyper-v), or |
| 58 | + QEMU/KVM for Linux). |
| 59 | + - No need to install a separate virtualization application. |
| 60 | + - Use of `cloud-init` to customize the VM. |
52 | 61 |
|
53 | 62 | ## Troubleshooting |
54 | 63 |
|
55 | | -For detailed troubleshooting instructions, please refer to the [troubleshooting |
56 | | -guide](/docs/other-guides/troubleshooting/). |
| 64 | +See the [Kubeflow troubleshooting guide](/docs/other-guides/troubleshooting/). |
57 | 65 |
|
58 | 66 | ## Resources |
59 | 67 |
|
60 | | -* The [documentation](/docs/) provides in-depth instructions for using Kubeflow |
| 68 | +* The [documentation](/docs/) provides in-depth instructions for using Kubeflow. |
61 | 69 | * Self-paced scenarios for learning and trying out Kubeflow: |
62 | 70 | * [Codelabs](https://codelabs.developers.google.com/?cat=tensorflow) |
63 | 71 | * [Introduction to Kubeflow on Google Kubernetes |
|
0 commit comments