{"id":546,"date":"2026-05-04T20:36:09","date_gmt":"2026-05-04T20:36:09","guid":{"rendered":"https:\/\/beginnerprojects.com\/cms\/?p=546"},"modified":"2026-05-04T20:36:09","modified_gmt":"2026-05-04T20:36:09","slug":"my-optimized-ollama-continue-setup-on-mac-studio","status":"publish","type":"post","link":"https:\/\/beginnerprojects.com\/cms\/my-optimized-ollama-continue-setup-on-mac-studio\/","title":{"rendered":"My Optimized Ollama + Continue Setup on Mac Studio"},"content":{"rendered":"\n<p>I was getting constant popup errors in Continue (VS Code extension) when using my Qwen3.6:27b model:&nbsp;<em>&#8220;Error parsing Ollama response: expected element type &lt;function&gt; but have &lt;parameter&gt;&#8221;<\/em>. Turns out it was a&nbsp;<strong>tool-calling template mismatch<\/strong>&nbsp;between Ollama and Continue, fixed by updating Ollama from 0.20.5 to 0.23.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"why-i-updated-ollama-and-how\">Why I Updated Ollama (and How)<\/h2>\n\n\n\n<p>I run Ollama on my Mac Studio M4 Max (36GB unified memory) and access it over LAN from my CachyOS laptop running VS Code + Continue + Cline extensions. The GUI makes updates dead simple:<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li>Click the Ollama menu bar icon<\/li>\n\n\n\n<li>Choose\u00a0<strong>&#8220;Restart to update&#8221;<\/strong><\/li>\n\n\n\n<li>Done! (<code>ollama --version<\/code>\u00a0confirmed 0.23)<\/li>\n<\/ol>\n\n\n\n<p><strong>Why it mattered<\/strong>: Ollama 0.23 includes updated&nbsp;<code>qwen3.5<\/code>&nbsp;renderer\/parser templates that fixed my Continue parsing errors completely. Older versions had mismatched tool-calling schemas.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"peeking-inside-ollama-show\">Peeking Inside:&nbsp;<code>ollama show<\/code><\/h2>\n\n\n\n<p>Curiosity led me to check what was actually happening with my model:<\/p>\n\n\n\n<pre class=\"wp-block-preformatted has-palette-color-7-background-color has-background\"><code>ollama show qwen3.6:27b --modelfile<\/code><\/pre>\n\n\n\n<p>Here&#8217;s the key section (ignoring license):<\/p>\n\n\n\n<pre class=\"wp-block-preformatted has-palette-color-7-background-color has-background\"><code>FROM \/Users\/me\/.ollama\/models\/blobs\/sha256-83c54730a5fea8a0958598c01617c1419c431e93b33bacf980b49a420c798926<br>TEMPLATE {{ .Prompt }}<br>RENDERER qwen3.5<br>PARSER qwen3.5<br>PARAMETER top_k 20<br>PARAMETER top_p 0.95<br>PARAMETER temperature 1<\/code><\/pre>\n\n\n\n<p><strong>What this tells me<\/strong>:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>No\u00a0<code>num_ctx<\/code><\/strong>\u00a0= uses global default (I set 32k in Ollama GUI)<\/li>\n\n\n\n<li><strong><code>RENDERER qwen3.5<\/code><\/strong>\u00a0= latest tool templates (fixed my errors)<\/li>\n\n\n\n<li><strong><code>top_k 20<\/code>,\u00a0<code>top_p 0.95<\/code>,\u00a0<code>temperature 1<\/code><\/strong>\u00a0= model defaults Continue was overriding<\/li>\n<\/ul>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"my-continue-config-that-actually-works\">My Continue Config That Actually Works<\/h2>\n\n\n\n<p>I access <a href=\"https:\/\/beginnerprojects.com\/cms\/run-your-own-ai-why-we-chose-ollama-for-local-intelligence\/\" data-type=\"post\" data-id=\"164\">Ollama<\/a> via\u00a0<code>OLLAMA_HOST=0.0.0.0 ollama serve<\/code>\u00a0on Mac, then connect from CachyOS over LAN. Here&#8217;s my\u00a0<code>~\/.continue\/config.yaml<\/code>\u00a0that&#8217;s optimized for my 36GB Mac Studio:<\/p>\n\n\n\n<pre class=\"wp-block-preformatted has-palette-color-7-background-color has-background\"><code>models:<br>  - name: qwen3.6:27b<br>    provider: ollama<br>    model: qwen3.6:27b<br>    apiBase: http:\/\/192.168.1.25:11434<br>    roles: [chat, edit, apply, summarize]<br>    capabilities: [tool_use]<br>    contextLength: 32768      # Matches my 32k GUI setting<br>    maxTokens: 4096<br>    timeout: 180000<br>    temperature: 0.2         # Override model's 1.0 (less random)<br>    top_p: 0.9<br>    top_k: 20                # Matches Modelfile (faster)<br><br>  - name: gemma4:31b<br>    provider: ollama<br>    model: gemma4:31b<br>    apiBase: http:\/\/192.168.1.25:11434<br>    roles: [chat, edit, apply, summarize]<br>    capabilities: [tool_use]<br>    contextLength: 32768<br>    maxTokens: 2048<br>    timeout: 180000<br>    temperature: 0.5<br>    top_p: 0.9<br>    top_k: 20<\/code><\/pre>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"why-these-specific-settings\">Why These Specific Settings<\/h2>\n\n\n\n<p><strong><code>contextLength: 32768<\/code><\/strong>: My Ollama GUI is set to 32k global default. Requesting 65k was wasting API negotiation time.<\/p>\n\n\n\n<p><strong><code>top_k: 20<\/code><\/strong>: Matches Modelfile exactly. Higher values (40) slow token generation 20-30%.<\/p>\n\n\n\n<p><strong><code>temperature: 0.2<\/code><\/strong>: Model default of\u00a0<code>1.0<\/code>\u00a0is too random for coding. 0.2 gives me a more focused responses.<\/p>\n\n\n\n<p><strong>36GB Mac Studio<\/strong>: Handles qwen3.6:27b at 32k context using ~28-32GB total. Smooth, no swapping.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"my-workflow\">My Workflow<\/h2>\n\n\n\n<pre class=\"wp-block-preformatted\"><code>Mac Studio (Ollama 0.23) \u2190LAN\u2192 CachyOS Laptop (VS Code + Continue + Cline)<br>                      \u2193<br>                 32k context, fast token generation<\/code><\/pre>\n\n\n\n<p>No more parsing popups, smooth 32k context, and Continue actually works reliably. The YouTubers show &#8220;ollama run model&#8221; demos, but for production workflows over LAN with Continue, these tweaks make all the difference.<\/p>\n\n\n\n<p><strong>Takeaway<\/strong>: Check your Modelfile, match your <a href=\"https:\/\/beginnerprojects.com\/cms\/the-secret-weapon-transforming-vscodium-with-the-continue-extension\/\" data-type=\"post\" data-id=\"167\">Continue<\/a> settings, update Ollama regularly. Your 30GB+ Mac can do way more than the defaults suggest.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>I was getting constant popup errors in Continue (VS Code extension) when using my Qwen3.6:27b model:&nbsp;&#8220;Error parsing Ollama response: expected element type &lt;function&gt; but have &lt;parameter&gt;&#8221;. Turns out it was a&nbsp;tool-calling template mismatch&nbsp;between Ollama and Continue, fixed by updating Ollama from 0.20.5 to 0.23. Why I Updated Ollama (and How) I run Ollama on my [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_ecmd_meta_description":"How I fixed Ollama parsing errors, tuned 32k context on Mac Studio 36GB, and optimized Continue config for CachyOS over LAN. Real settings that actually work.","footnotes":""},"categories":[1],"tags":[],"class_list":["post-546","post","type-post","status-publish","format-standard","hentry","category-guides"],"blocksy_meta":[],"_links":{"self":[{"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/posts\/546","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/comments?post=546"}],"version-history":[{"count":1,"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/posts\/546\/revisions"}],"predecessor-version":[{"id":547,"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/posts\/546\/revisions\/547"}],"wp:attachment":[{"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/media?parent=546"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/categories?post=546"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/beginnerprojects.com\/cms\/wp-json\/wp\/v2\/tags?post=546"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}