https://clarence.tw/en/notes/llm/ Recent content in LLM & AI on Clarence | Backend, AI, DevOps, Infrastructure Hugo -- gohugo.io en https://clarence.tw/en/notes/llm/ollama-setup/ Mon, 01 Jan 0001 00:00:00 +0000 https://clarence.tw/en/notes/llm/ollama-setup/ <div class="note-card "> <div class="item"> <h5 class="note-title"><span>Install and run</span></h5> <div class="card"> <div class="card-body"><div class="highlight"><pre tabindex="0" style="color:#f8f8f2;background-color:#272822;-moz-tab-size:4;-o-tab-size:4;tab-size:4;-webkit-text-size-adjust:none;"><code class="language-bash" data-lang="bash"><span style="display:flex;"><span>curl -fsSL https://ollama.com/install.sh | sh </span></span><span style="display:flex;"><span>ollama serve </span></span><span style="display:flex;"><span>ollama pull gemma2:9b </span></span><span style="display:flex;"><span>ollama pull llama3.1:8b </span></span><span style="display:flex;"><span>ollama list </span></span></code></pre></div><p>Use smaller models first when testing prompts, then move to larger models after the workflow is stable.</p></div> </div> </div> </div> <div class="note-card "> <div class="item"> <h5 class="note-title"><span>Model commands</span></h5> <div class="card"> <div class="card-body"><div class="highlight"><pre tabindex="0" style="color:#f8f8f2;background-color:#272822;-moz-tab-size:4;-o-tab-size:4;tab-size:4;-webkit-text-size-adjust:none;"><code class="language-bash" data-lang="bash"><span style="display:flex;"><span>ollama run gemma2:9b </span></span><span style="display:flex;"><span>ollama show gemma2:9b </span></span><span style="display:flex;"><span>ollama rm gemma2:9b </span></span></code></pre></div><p>Keep model names explicit in scripts so a future default model change does not alter results silently.</p></div> </div> </div> </div> <div class="note-card "> <div class="item"> <h5 class="note-title"><span>Local API</span></h5> <div class="card"> <div class="card-body"><div class="highlight"><pre tabindex="0" style="color:#f8f8f2;background-color:#272822;-moz-tab-size:4;-o-tab-size:4;tab-size:4;-webkit-text-size-adjust:none;"><code class="language-bash" data-lang="bash"><span style="display:flex;"><span>curl http://localhost:11434/api/generate <span style="color:#ae81ff">\ </span></span></span><span style="display:flex;"><span> -d <span style="color:#e6db74">&#39;{&#34;model&#34;:&#34;gemma2:9b&#34;,&#34;prompt&#34;:&#34;Summarize this text.&#34;,&#34;stream&#34;:false}&#39;</span> </span></span></code></pre></div><p>Local API calls are useful for private experiments, but production systems still need timeout, retry, and logging boundaries.</p>