Deploying to gh-pages from @ dstackai/dstack@61eb095 🚀

Author: peterschmidt85

blog/amd-mi300x-inference-benchmark/index.html

@@ -92,7 +92,7 @@
 <meta property="og:title" content="Benchmarking Llama 3.1 405B on 8x AMD MI300X GPUs - dstack" />
 <meta property="og:description" content="Exploring how the inference performance of Llama 3.1 405B varies on 8x AMD MI300X GPUs across vLLM and TGI backends in different use cases." />
 <meta property="og:image" content="
-  https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-hotaisle-amd-mi300x-prompt-v5.png?raw=true" />
+  https://dstack.ai/static-assets/static-assets/images/dstack-hotaisle-amd-mi300x-prompt-v5.png" />
 <meta property="og:image:type" content="image/png" />
 <meta property="og:image:width" content="1200" />
 <meta property="og:image:height" content="630" />
@@ -101,7 +101,7 @@
 <meta property="twitter.title" content="Benchmarking Llama 3.1 405B on 8x AMD MI300X GPUs - dstack" />
 <meta property="twitter:description" content="Exploring how the inference performance of Llama 3.1 405B varies on 8x AMD MI300X GPUs across vLLM and TGI backends in different use cases." />
 <meta property="twitter:image" content="
-  https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-hotaisle-amd-mi300x-prompt-v5.png?raw=true" />
+  https://dstack.ai/static-assets/static-assets/images/dstack-hotaisle-amd-mi300x-prompt-v5.png" />
 </head>
 
 
@@ -3903,7 +3903,7 @@ <h1 id="benchmarking-llama-31-405b-on-8x-amd-mi300x-gpus">Benchmarking Llama 3.1
 so we saw this as a great chance to test our integration by benchmarking AMD GPUs. Our friends at 
 <a href="https://hotaisle.xyz/" target="_blank">Hot Aisle <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a>, who build top-tier 
 bare metal compute for AMD GPUs, kindly provided the hardware for the benchmark.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-hotaisle-amd-mi300x-prompt-v5.png?raw=true" width="750" /></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-hotaisle-amd-mi300x-prompt-v5.png" width="750" /></p>
 <!-- more -->
 
 <p>With access to a bare metal machine with 8x AMD MI300X GPUs from Hot Aisle, we decided to skip smaller models and went

blog/amd-on-runpod/index.html

@@ -3850,7 +3850,7 @@ <h2 id="configuration">Configuration<a class="headerlink" href="#configuration"
 <summary>Control plane</summary>
 <p>If you specify <code>model</code> when running a service, <code>dstack</code> will automatically register the model on
 an OpenAI-compatible endpoint and allow you to use it for chat via the control plane UI.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-control-plane-model-llama31.png?raw=true" width="750px" /></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-control-plane-model-llama31.png" width="750px" /></p>
 </details>
 <h2 id="whats-next">What's next?<a class="headerlink" href="#whats-next" title="Permanent link">&para;</a></h2>
 <ol>

blog/amd-on-tensorwave/index.html

@@ -92,7 +92,7 @@
 <meta property="og:title" content="Using SSH fleets with TensorWave's private AMD cloud - dstack" />
 <meta property="og:description" content="This tutorial walks you through how dstack can be used with TensorWave's private AMD cloud using SSH fleets." />
 <meta property="og:image" content="
-  https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-tensorwave-v2.png?raw=true" />
+  https://dstack.ai/static-assets/static-assets/images/dstack-tensorwave-v2.png" />
 <meta property="og:image:type" content="image/png" />
 <meta property="og:image:width" content="1200" />
 <meta property="og:image:height" content="630" />
@@ -101,7 +101,7 @@
 <meta property="twitter.title" content="Using SSH fleets with TensorWave's private AMD cloud - dstack" />
 <meta property="twitter:description" content="This tutorial walks you through how dstack can be used with TensorWave's private AMD cloud using SSH fleets." />
 <meta property="twitter:image" content="
-  https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-tensorwave-v2.png?raw=true" />
+  https://dstack.ai/static-assets/static-assets/images/dstack-tensorwave-v2.png" />
 </head>
 
 
@@ -3788,14 +3788,14 @@ <h1 id="using-ssh-fleets-with-tensorwaves-private-amd-cloud">Using SSH fleets wi
 <p>In this tutorial, we’ll walk you through how <code>dstack</code> can be used with
 <a href="https://tensorwave.com/" target="_blank">TensorWave <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a> using 
 <a href="../../docs/concepts/fleets/#ssh">SSH fleets</a>.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-tensorwave-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-tensorwave-v2.png" width="630"/></p>
 <!-- more -->
 
 <p>TensorWave is a cloud provider specializing in large-scale AMD GPU clusters for both
 training and inference.</p>
 <p>Before following this tutorial, ensure you have access to a cluster. You’ll see the cluster and its nodes in your
 TensorWave dashboard.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-tensorwave-ui.png?raw=true" width="750"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-tensorwave-ui.png" width="750"/></p>
 <h2 id="creating-a-fleet">Creating a fleet<a class="headerlink" href="#creating-a-fleet" title="Permanent link">&para;</a></h2>
 <details class="info">
 <summary>Prerequisites</summary>

blog/archive/2024/index.html

@@ -3969,7 +3969,7 @@ <h2 id="exploring-inference-memory-saturation-effect-h100-vs-mi300x"><a class="t
 Additionally, we compare deployment strategies: running two Llama 3.1 405B FP8 replicas on 4xMI300x versus a single
 replica on 4xMI300x and 8xMI300x</p>
 <p>Finally, we extrapolate performance projections for upcoming GPUs like NVIDIA H200, B200, and AMD MI325x, MI350x.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/h100-mi300x-inference-benchmark-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/h100-mi300x-inference-benchmark-v2.png" width="630"/></p>
 <p>This benchmark is made possible through the generous support of our friends at
 <a href="https://hotaisle.xyz/" target="_blank">Hot Aisle <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a> and 
 <a href="https://lambdalabs.com/" target="_blank">Lambda <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a>,
@@ -4147,7 +4147,7 @@ <h3 id="how-it-works" style="display:none"><a class="toclink" href="../../dstack
 <p>While it's possible to use third-party monitoring tools with <code>dstack</code>, it is often more convenient to debug your run and
 track metrics out of the box. That's why, with the latest release, <code>dstack</code> introduced <a href="../../../docs/reference/cli/dstack/metrics/"><code>dstack stats</code></a>, a new CLI (and API)
 for monitoring container metrics, including GPU usage for <code>NVIDIA</code>, <code>AMD</code>, and other accelerators.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-stats-v2.png?raw=true" width="725"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-stats-v2.png" width="725"/></p>
 
 
       <nav class="md-post__action">
@@ -4194,7 +4194,7 @@ <h2 id="benchmarking-llama-31-405b-on-8x-amd-mi300x-gpus"><a class="toclink" hre
 so we saw this as a great chance to test our integration by benchmarking AMD GPUs. Our friends at 
 <a href="https://hotaisle.xyz/" target="_blank">Hot Aisle <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a>, who build top-tier 
 bare metal compute for AMD GPUs, kindly provided the hardware for the benchmark.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-hotaisle-amd-mi300x-prompt-v5.png?raw=true" width="750" /></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-hotaisle-amd-mi300x-prompt-v5.png" width="750" /></p>
 
 
       <nav class="md-post__action">

blog/archive/2025/index.html

@@ -3919,7 +3919,7 @@ <h2 id="supporting-gpu-provisioning-and-orchestration-on-nebius"><a class="tocli
 developer velocity and efficiency.
 <code>dstack</code> is an open-source orchestrator purpose-built for AI infrastructure—offering a lightweight, container-native
 alternative to Kubernetes and Slurm.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-nebius-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-nebius-v2.png" width="630"/></p>
 <p>Today, we’re announcing native integration with <a href="https://nebius.com/" target="_blank">Nebius <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a>,
 offering a streamlined developer experience for teams using GPUs for AI workloads.</p>
 
@@ -3968,7 +3968,7 @@ <h2 id="built-in-ui-for-monitoring-essential-gpu-metrics"><a class="toclink" hre
 <p>AI workloads generate vast amounts of metrics, making it essential to have efficient monitoring tools. While our recent
 update introduced the ability to export available metrics to Prometheus for maximum flexibility, there are times when
 users need to quickly access essential metrics without the need to switch to an external tool.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-metrics-ui-v3-min.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-metrics-ui-v3-min.png" width="630"/></p>
 <p>Previously, we introduced a <a href="../../dstack-metrics/">CLI command</a> that allows users to view essential GPU metrics for both NVIDIA
 and AMD hardware. Now, with this latest update, we’re excited to announce the addition of a built-in dashboard within
 the <code>dstack</code> control plane.</p>
@@ -4022,7 +4022,7 @@ <h2 id="supporting-mpi-and-ncclrccl-tests"><a class="toclink" href="../../mpi/">
 <code>torchrun</code>, <code>accelerate</code>, or others. <code>dstack</code> handles node provisioning, job execution, and automatically propagates
 system environment variables—such as <code>DSTACK_NODE_RANK</code>, <code>DSTACK_MASTER_NODE_IP</code>,
 <code>DSTACK_GPUS_PER_NODE</code> and <a href="../../../docs/concepts/tasks/#system-environment-variables">others</a>—to containers.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-mpi-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-mpi-v2.png" width="630"/></p>
 <p>One use case <code>dstack</code> hasn’t supported until now is MPI, as it requires a scheduled environment or
 direct SSH connections between containers. Since <code>mpirun</code> is essential for running NCCL/RCCL tests—crucial for large-scale
 cluster usage—we’ve added support for it.</p>
@@ -4075,7 +4075,7 @@ <h3 id="why-prometheus" style="display:none"><a class="toclink" href="../../prom
 <p>While <code>dstack</code> provides key metrics through its UI and <a href="../../dstack-metrics/"><code>dstack metrics</code></a> CLI, teams often need more granular data and prefer
 using their own monitoring tools. To support this, we’ve introduced a new endpoint that allows real-time exporting all collected
 metrics—covering fleets and runs—directly to Prometheus.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-prometheus-v3.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-prometheus-v3.png" width="630"/></p>
 
 
       <nav class="md-post__action">
@@ -4122,7 +4122,7 @@ <h2 id="accessing-dev-environments-with-cursor"><a class="toclink" href="../../c
 <p>Previously, support was limited to VS Code. However, as developers rely on a variety of desktop IDEs,
 we’ve expanded compatibility. With this update, dev environments now offer effortless access for users of 
 <a href="https://www.cursor.com/" target="_blank">Cursor <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a>.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-cursor-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-cursor-v2.png" width="630"/></p>
 
 
       <nav class="md-post__action">
@@ -4172,7 +4172,7 @@ <h2 id="deepseek-r1-inference-performance-mi300x-vs-h200"><a class="toclink" hre
 <p>In this benchmark, we evaluate the performance of three inference backends—SGLang, vLLM, and TensorRT-LLM—on two hardware
 configurations: 8x NVIDIA H200 and 8x AMD MI300X. Our goal is to compare throughput, latency, and overall efficiency to
 determine the optimal backend and hardware pairing for DeepSeek-R1's demanding requirements.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/h200-mi300x-deepskeek-benchmark-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/h200-mi300x-deepskeek-benchmark-v2.png" width="630"/></p>
 <p>This benchmark was made possible through the generous support of our partners at
 <a href="https://www.vultr.com/" target="_blank">Vultr <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a> and 
 <a href="https://lambdalabs.com/" target="_blank">Lambda <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a>,
@@ -4224,7 +4224,7 @@ <h2 id="using-ssh-fleets-with-tensorwaves-private-amd-cloud"><a class="toclink"
 <p>In this tutorial, we’ll walk you through how <code>dstack</code> can be used with
 <a href="https://tensorwave.com/" target="_blank">TensorWave <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a> using 
 <a href="../../../docs/concepts/fleets/#ssh">SSH fleets</a>.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-tensorwave-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-tensorwave-v2.png" width="630"/></p>
 
 
       <nav class="md-post__action">
@@ -4271,7 +4271,7 @@ <h2 id="supporting-intel-gaudi-ai-accelerators-with-ssh-fleets"><a class="toclin
 just leading cloud providers and on-prem environments but also a wide range of accelerators.</p>
 <p>With our latest release, we’re adding support
 for Intel Gaudi AI Accelerator and launching a new partnership with Intel.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-intel-gaudi-and-intel-tiber-cloud-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-intel-gaudi-and-intel-tiber-cloud-v2.png" width="630"/></p>
 
 
       <nav class="md-post__action">
@@ -4317,7 +4317,7 @@ <h2 id="efficient-distributed-training-with-aws-efa"><a class="toclink" href="..
 ultra-low latency and high-throughput communication between nodes. This makes it an ideal solution for scaling
 distributed training workloads across multiple GPUs and instances.</p>
 <p>With the latest release of <code>dstack</code>, you can now leverage AWS EFA to supercharge your distributed training tasks.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/distributed-training-with-aws-efa-v2.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/distributed-training-with-aws-efa-v2.png" width="630"/></p>
 
 
       <nav class="md-post__action">
@@ -4365,7 +4365,7 @@ <h2 id="auto-shutdown-for-inactive-dev-environmentsno-idle-gpus"><a class="tocli
 a container that has GPU access.</p>
 <p>One issue with dev environments is forgetting to stop them or closing your laptop, leaving the GPU idle and costly. With
 our latest update, <code>dstack</code> now detects inactive environments and automatically shuts them down, saving you money.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/inactive-dev-environments-auto-shutdown.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/inactive-dev-environments-auto-shutdown.png" width="630"/></p>
 
 
       <nav class="md-post__action">

blog/archive/2025/page/2/index.html

@@ -3744,7 +3744,7 @@ <h2 id="introducing-gpu-blocks-and-proxy-jump-for-ssh-fleets"><a class="toclink"
 <p>Originally, <code>dstack</code> was focused on public clouds. With the new release, <code>dstack</code>
 extends support to data centers and private clouds, offering a simpler, AI-native solution that replaces Kubernetes and
 Slurm.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/data-centers-and-private-clouds.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/data-centers-and-private-clouds.png" width="630"/></p>
 
 
       <nav class="md-post__action">
@@ -3794,7 +3794,7 @@ <h2 id="supporting-nvidia-and-amd-accelerators-on-vultr"><a class="toclink" href
 approach.
 Today, we’re excited to share a new integration and partnership
 with <a href="https://www.vultr.com/" target="_blank">Vultr <span class="twemoji external"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.93 5 2.83 2.83L5 17.59 6.42 19l9.76-9.75L19 12.07V5z"/></svg></span></a>.</p>
-<p><img src="https://github.com/dstackai/static-assets/blob/main/static-assets/images/dstack-vultr.png?raw=true" width="630"/></p>
+<p><img src="https://dstack.ai/static-assets/static-assets/images/dstack-vultr.png" width="630"/></p>
 <p>This new integration enables Vultr customers to train and deploy models on both AMD
 and NVIDIA GPUs with greater flexibility and efficiency–using <code>dstack</code>. </p>