Deploying to gh-pages from @ dstackai/dstack@1d931dc 🚀

Author: peterschmidt85

assets/images/social/blog/archive/2024.png

@@ -2172,6 +2172,17 @@
     </label>
     <ul class="md-nav__list" data-md-component="toc" data-md-scrollfix>
 
+        <li class="md-nav__item">
+  <a href="#optimizing-inference-cold-starts-on-runpod-with-volumes" class="md-nav__link">
+    <span class="md-ellipsis">
+      
+        Optimizing inference cold starts on RunPod with volumes
+      
+    </span>
+  </a>
+  
+</li>
+      
         <li class="md-nav__item">
   <a href="#dstack-sky-now-allows-using-your-own-cloud-accounts" class="md-nav__link">
     <span class="md-ellipsis">
@@ -2322,6 +2333,17 @@
     </label>
     <ul class="md-nav__list" data-md-component="toc" data-md-scrollfix>
 
+        <li class="md-nav__item">
+  <a href="#optimizing-inference-cold-starts-on-runpod-with-volumes" class="md-nav__link">
+    <span class="md-ellipsis">
+      
+        Optimizing inference cold starts on RunPod with volumes
+      
+    </span>
+  </a>
+  
+</li>
+      
         <li class="md-nav__item">
   <a href="#dstack-sky-now-allows-using-your-own-cloud-accounts" class="md-nav__link">
     <span class="md-ellipsis">
@@ -2362,6 +2384,46 @@ <h1 id="2024">2024<a class="headerlink" href="#2024" title="Permanent link">&par
         <article class="md-post md-post--excerpt">
   <header class="md-post__header">
 
+    <div class="md-post__meta md-meta">
+      <ul class="md-meta__list">
+        <li class="md-meta__item">
+          <time datetime="2024-08-13 00:00:00">August 13, 2024</time></li>
+        
+        
+          
+          <li class="md-meta__item">
+            
+              2 min read
+            
+          </li>
+        
+        
+      </ul>
+      
+    </div>
+  </header>
+  <div class="md-post__content md-typeset">
+    <h2 id="optimizing-inference-cold-starts-on-runpod-with-volumes"><a class="toclink" href="../../volumes-on-runpod/">Optimizing inference cold starts on RunPod with volumes</a></h2>
+<p>Deploying custom models in the cloud often faces the challenge of cold start times, including the time to provision a
+new instance and download the model. This is especially relevant for services with autoscaling when new model replicas
+need to be provisioned quickly. </p>
+<p>Let's explore how <code>dstack</code> optimizes this process using volumes, with an example of
+deploying a model on RunPod.</p>
+
+    
+      <nav class="md-post__action">
+        <a href="../../volumes-on-runpod/">
+          Continue reading
+        </a>
+      </nav>
+    
+    
+  </div>
+</article>
+      
+        <article class="md-post md-post--excerpt">
+  <header class="md-post__header">
+    
     <div class="md-post__meta md-meta">
       <ul class="md-meta__list">
         <li class="md-meta__item">

assets/images/social/blog/index.png

@@ -18,6 +18,8 @@
         <link rel="prev" href="../dstack-sky/">
 
 
+        <link rel="next" href="../volumes-on-runpod/">
+      
 
 
 

assets/images/social/blog/volumes-on-runpod.png

@@ -2088,6 +2088,17 @@
     </label>
     <ul class="md-nav__list" data-md-component="toc" data-md-scrollfix>
 
+        <li class="md-nav__item">
+  <a href="#optimizing-inference-cold-starts-on-runpod-with-volumes" class="md-nav__link">
+    <span class="md-ellipsis">
+      
+        Optimizing inference cold starts on RunPod with volumes
+      
+    </span>
+  </a>
+  
+</li>
+      
         <li class="md-nav__item">
   <a href="#dstack-sky-now-allows-using-your-own-cloud-accounts" class="md-nav__link">
     <span class="md-ellipsis">
@@ -2322,6 +2333,17 @@
     </label>
     <ul class="md-nav__list" data-md-component="toc" data-md-scrollfix>
 
+        <li class="md-nav__item">
+  <a href="#optimizing-inference-cold-starts-on-runpod-with-volumes" class="md-nav__link">
+    <span class="md-ellipsis">
+      
+        Optimizing inference cold starts on RunPod with volumes
+      
+    </span>
+  </a>
+  
+</li>
+      
         <li class="md-nav__item">
   <a href="#dstack-sky-now-allows-using-your-own-cloud-accounts" class="md-nav__link">
     <span class="md-ellipsis">
@@ -2362,6 +2384,46 @@ <h1 id="blog">Blog<a class="headerlink" href="#blog" title="Permanent link">&par
         <article class="md-post md-post--excerpt">
   <header class="md-post__header">
 
+    <div class="md-post__meta md-meta">
+      <ul class="md-meta__list">
+        <li class="md-meta__item">
+          <time datetime="2024-08-13 00:00:00">August 13, 2024</time></li>
+        
+        
+          
+          <li class="md-meta__item">
+            
+              2 min read
+            
+          </li>
+        
+        
+      </ul>
+      
+    </div>
+  </header>
+  <div class="md-post__content md-typeset">
+    <h2 id="optimizing-inference-cold-starts-on-runpod-with-volumes"><a class="toclink" href="volumes-on-runpod/">Optimizing inference cold starts on RunPod with volumes</a></h2>
+<p>Deploying custom models in the cloud often faces the challenge of cold start times, including the time to provision a
+new instance and download the model. This is especially relevant for services with autoscaling when new model replicas
+need to be provisioned quickly. </p>
+<p>Let's explore how <code>dstack</code> optimizes this process using volumes, with an example of
+deploying a model on RunPod.</p>
+
+    
+      <nav class="md-post__action">
+        <a href="volumes-on-runpod/">
+          Continue reading
+        </a>
+      </nav>
+    
+    
+  </div>
+</article>
+      
+        <article class="md-post md-post--excerpt">
+  <header class="md-post__header">
+    
     <div class="md-post__meta md-meta">
       <ul class="md-meta__list">
         <li class="md-meta__item">