https://deardhruv.com/tags/diffusion-models/Recent content in Diffusion Models on Dhruv Patel - Lead Android EngineerHugoendhruv.time@gmail.com (Dhruv Patel)dhruv.time@gmail.com (Dhruv Patel)<a href="https://creativecommons.org/licenses/by-nc/4.0/" target="_blank" rel="noopener">CC BY-NC 4.0</a>Mon, 05 Oct 2026 16:26:02 +0530https://deardhruv.com/posts/ai_playground_series/04-i-put-an-image-generator-inside-the-phone/Mon, 05 Oct 2026 13:00:00 +0530dhruv.time@gmail.com (Dhruv Patel)https://deardhruv.com/posts/ai_playground_series/04-i-put-an-image-generator-inside-the-phone/<h1 id="i-put-an-image-generator-inside-the-phone">I Put an Image Generator Inside the Phone</h1> <h2 id="part-4---i-did-not-build-a-diffusion-engine-i-integrated-image-generation-runtimes-and-then-had-to-engineer-everything-around-them">Part 4 - I did not build a diffusion engine. I integrated image-generation runtimes and then had to engineer everything around them.</h2> <p>Text generation had already forced me to learn about:</p> <div class="highlight"><pre tabindex="0" class="chroma"><code class="language-text" data-lang="text"><span class="line"><span class="cl">runtime selection </span></span><span class="line"><span class="cl">memory </span></span><span class="line"><span class="cl">native code </span></span><span class="line"><span class="cl">lifecycle </span></span><span class="line"><span class="cl">fallback </span></span></code></pre></div><p>Image generation brought all of those problems back, but in a different shape.</p> <p>The important distinction at the start is the same one I made in Part 1:</p>