You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
<p><code>dstack</code> supports <code>NVIDIA</code>, <code>AMD</code>, <code>Google TPU</code>, and <code>Intel Gaudi</code> accelerators out of the box.</p>
3675
3649
<h2id="how-does-it-work">How does it work?<aclass="headerlink" href="#how-does-it-work" title="Permanent link">¶</a></h2>
<h4id="1-set-up-the-server">1. Set up the server<aclass="headerlink" href="#1-set-up-the-server" title="Permanent link">¶</a></h4>
3678
3652
<blockquote>
3679
3653
<p>Before using <code>dstack</code>, ensure you've <ahref="installation/">installed</a> the server, or signed up for <ahref="https://sky.dstack.ai" target="_blank">dstack Sky <spanclass="twemoji external"><svgxmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><pathd="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a>.</p>
<p><code>dstack</code> streamlines infrastructure management and container usage, enabling AI teams to work with any frameworks across
3701
3675
cloud platforms or on-premise servers.</p>
3702
-
<h2id="how-does-it-compare-to-other-tools">How does it compare to other tools?<aclass="headerlink" href="#how-does-it-compare-to-other-tools" title="Permanent link">¶</a></h2>
3703
-
<detailsclass="info">
3704
-
<summary>How does dstack compare to Kubernetes?</summary>
3705
-
<p><code>dstack</code> is more lightweight, and is designed specifically for AI, enabling AI engineers to handle development, training, and
3706
-
deployment without needing extra tools or Ops support. </p>
3707
-
<p>With <code>dstack</code>, you don't need Kubeflow or other ML platforms on top—everything is available out of the box.</p>
3708
-
<p>Additionally, <code>dstack</code> is much easier to use with on-prem servers—just provide hostnames and SSH credentials,
3709
-
and <code>dstack</code> will automatically create a fleet ready for use with development environments, tasks, and services.</p>
3710
-
</details>
3711
-
<detailsclass="info">
3712
-
<summary>Can dstack and Kubernetes be used together?</summary>
3713
-
<p>For AI development, it’s more efficient to use <code>dstack</code> directly with your cloud accounts or on-prem servers—without Kubernetes.</p>
3714
-
<p>However, if you prefer, you can set up the <code>dstack</code> server with a Kubernetes backend to provision through Kubernetes.</p>
3715
-
<p>Does your Ops team insist on using Kubernetes for production-grade deployment? You can use <code>dstack</code> and
3716
-
Kubernetes side by side; <code>dstack</code> for development and Kubernetes for production-grade deployment.</p>
3717
-
</details>
3718
-
<detailsclass="info">
3719
-
<summary>How does dstack compare to KubeFlow?</summary>
3720
-
<p><code>dstack</code> can be used entirely instead of Kubeflow. It covers everything that Kubeflow does, and much more on top,
3721
-
including development environments, services, and additional features.</p>
3722
-
<p><code>dstack</code> is easier to set up with on-premises servers, doesn't require Kubernetes, and works with multiple cloud
3723
-
providers out of the box.</p>
3724
-
</details>
3725
-
<detailsclass="info">
3726
-
<summary>How does dstack compare to Slurm?</summary>
3727
-
<p><code>dstack</code> can be used entirely instead of Slurm. It covers everything that Slurm does, and a lot more on top, including
3728
-
dev environments, services, out-of-the-box cloud support, easier setup with on-premises servers, and much more.</p>
0 commit comments