<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Anibal Guerrero Hernandez on Medium]]></title>
        <description><![CDATA[Stories by Anibal Guerrero Hernandez on Medium]]></description>
        <link>https://medium.com/@anibal.guerrero.hernandez?source=rss-a5c18f515288------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*fevtpZrBzYeIXYuP</url>
            <title>Stories by Anibal Guerrero Hernandez on Medium</title>
            <link>https://medium.com/@anibal.guerrero.hernandez?source=rss-a5c18f515288------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 20:02:58 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@anibal.guerrero.hernandez/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[ Unlocking HiDream on macOS: A New Open-Source AI Model with ComfyUI]]></title>
            <link>https://medium.com/@anibal.guerrero.hernandez/unlocking-hidream-on-macos-a-new-open-source-ai-model-with-comfyui-f97abdeea4aa?source=rss-a5c18f515288------2</link>
            <guid isPermaLink="false">https://medium.com/p/f97abdeea4aa</guid>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[image-generation]]></category>
            <category><![CDATA[apple-silicon]]></category>
            <category><![CDATA[comfy-ui]]></category>
            <category><![CDATA[hidream]]></category>
            <dc:creator><![CDATA[Anibal Guerrero Hernandez]]></dc:creator>
            <pubDate>Tue, 22 Apr 2025 18:54:24 GMT</pubDate>
            <atom:updated>2025-04-22T18:54:24.101Z</atom:updated>
            <content:encoded><![CDATA[<p>The AI community has a brand new toy to play with: <strong>HiDream</strong>, an open-source generative model with impressive capabilities. Thanks to its integration into <strong>ComfyUI</strong>, you can run this model locally and explore its potential right now. This tutorial will guide you through the process, especially if you’re on Apple Silicon, where things require a few extra steps.</p><h3>🚀 What is HiDream?</h3><p>HiDream is a newly released, cutting-edge generative model that delivers high-quality image outputs. It supports various levels of fidelity and performance, including Fast, Dev, and Full pipelines, making it accessible whether you’re experimenting on a laptop or running powerful GPUs.</p><p>Best of all? It’s <strong>open source</strong> and now <strong>natively supported by ComfyUI</strong>.</p><h3>💡 Why Use ComfyUI for HiDream?</h3><p>ComfyUI is a modular, node-based interface for running AI workflows locally. It now supports HiDream templates out of the box, making it incredibly easy to generate images using this powerful model.</p><p>Whether you want to tinker with individual nodes or run full templates, ComfyUI gives you a visual playground for AI creativity.</p><h3>✅ Step 1: Install ComfyUI</h3><p>First, set up ComfyUI. Follow this full setup guide if you haven’t already:</p><p>👉 <a href="https://medium.com/@anibal.guerrero.hernandez/how-to-install-comfyui-on-macos-m1-m2-m3-or-m4-a01f810a0ca3">How to Install ComfyUI Locally</a></p><h3>⚖️ Step 2: Get HiDream Running (Apple Silicon Users Read This!)</h3><p>HiDream depends on <strong>Triton</strong>, which doesn’t install via pip on Apple Silicon. You need to build it from source:</p><pre>git clone https://github.com/triton-lang/triton.git<br>cd triton</pre><pre>pip install -r python/requirements.txt<br>pip install -e python</pre><blockquote><em>🔧 This step is mandatory for macOS users. Skip it and HiDream won’t work.</em></blockquote><h3>📂 Step 3: Add HiDream to ComfyUI</h3><p>Clone the HiDream Sampler into your custom_nodes folder:</p><pre>git clone https://github.com/lum3on/comfyui_HiDream-Sampler ComfyUI/custom_nodes/comfyui_HiDream-Sampler</pre><p>Install the necessary Python packages:</p><pre>python -m pip install -r custom_nodes/comfyui_HiDream-Sampler/requirements.txt --no-build-isolation</pre><p>Then, <strong>restart ComfyUI</strong>.</p><h3>⚡ Troubleshooting: Float8_e4m3fn Errors (Apple Silicon)</h3><p>Trying to run main.py directly might throw this error:</p><pre>Trying to convert Float8_e4m3fn to the MPS backend but it does not have support for that dtype</pre><p>Fix it by launching ComfyUI with these flags:</p><pre>python main.py --force-fp16 --use-split-cross-attention --cpu</pre><h3>🔹 Step 4: Open a HiDream Template</h3><p>Inside ComfyUI, go to <strong>Workflow → Browse Templates</strong>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/724/1*tq3hxIH--QanJefvTS3rfQ.png" /><figcaption>Browse through the workflow options into the templates.</figcaption></figure><p>Then, go to<strong> Templates → Image → HiDream</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ONkRLerVIOgLuHNziSBDAQ.png" /><figcaption>Templates included in the Image Generation section, with different HiDream models available.</figcaption></figure><p>Choose from:</p><ul><li><strong>Fast - </strong>Good for quick experiments</li><li><strong>Dev - </strong>Our recommended sweet spot ✅</li><li><strong>Full - </strong>Best quality, most resource-heavy</li></ul><h3>📊 Download the Required Model Files</h3><p>HiDream needs a handful of .safetensors models. ComfyUI will tell you which ones are missing. You can download them from inside the app.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AVSKKJ8JQOyeAzg2l5-Osw.png" /><figcaption>Warning with missing nodes required.</figcaption></figure><p>You’ll need ~32 GB of space. Here are the models:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lum4Kj2UjGQSA4r58s1z5g.png" /><figcaption>Missing models and their corresponding folder locations and sizes.</figcaption></figure><p>Once downloaded, reopen the template. It should look like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lH9ZGUzSqb0EUgxfZKWmqw.png" /><figcaption>Template workflow loaded for HiDream I1 Dev.</figcaption></figure><h3>🎬 Run Your First HiDream Generation</h3><p>Click <strong>Run</strong> in ComfyUI and watch HiDream do its thing. You should see a beautiful image generated based on your workflow settings.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sHsx1h3mrjSrtDkHD1XNAw.png" /><figcaption>Output image obtained from running sample prompt with HiDream I1.</figcaption></figure><h3>💭 Final Thoughts &amp; What I’m Building</h3><p>HiDream is one of the most exciting open-source model releases in the generative AI space, and ComfyUI makes it surprisingly easy to experiment with.</p><p>Whether you’re an artist, developer, or AI enthusiast, this is your chance to push the boundaries of what’s possible <strong>locally</strong> on macOS.</p><p>If you found this guide helpful, feel free to share it and connect with me over on X:<br><strong>@</strong><a href="https://x.com/fortranibal"><strong>fortranibal</strong></a> 🧠✨</p><blockquote><em>I’m also building </em><a href="https://www.tryreflektion.com/"><strong><em>Reflektion</em></strong></a><em>, an AI therapy assistant designed for meaningful conversations, emotional continuity, and deep personalization. Affordable, accessible, and always there for you.</em></blockquote><blockquote><strong><em>Find your path with Reflektion.</em></strong></blockquote><p>Questions? Thoughts? Share your workflows and results in the comments!</p><p>Happy creating! 🎨🤖</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f97abdeea4aa" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ How to Install ComfyUI on MacOS (M1, M2, M3, or M4)]]></title>
            <link>https://medium.com/@anibal.guerrero.hernandez/how-to-install-comfyui-on-macos-m1-m2-m3-or-m4-a01f810a0ca3?source=rss-a5c18f515288------2</link>
            <guid isPermaLink="false">https://medium.com/p/a01f810a0ca3</guid>
            <category><![CDATA[mac]]></category>
            <category><![CDATA[local-deployment]]></category>
            <category><![CDATA[video-generation]]></category>
            <category><![CDATA[image-generation]]></category>
            <category><![CDATA[comfy-ui]]></category>
            <dc:creator><![CDATA[Anibal Guerrero Hernandez]]></dc:creator>
            <pubDate>Mon, 21 Apr 2025 17:50:38 GMT</pubDate>
            <atom:updated>2025-04-21T17:50:38.501Z</atom:updated>
            <content:encoded><![CDATA[<p>If you’re ready to unlock the full potential of <strong>ComfyUI</strong> on your M1, M2, M3, or M4 Mac, this step-by-step guide has you covered. <strong>ComfyUI</strong> is a modular, node-based interface that makes it easy to build and experiment with powerful AI workflows. Whether you’re working with <strong>Stable Diffusion</strong>, <strong>HiDream</strong>, <strong>Flux</strong>, or other cutting-edge models, ComfyUI is an ideal platform for generating AI-powered images, animations, and more.</p><blockquote><em>✅ This guide covers setup, installation of dependencies, and how to launch ComfyUI on your Mac.</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Uq8hyp61eDztbck2.jpg" /><figcaption>Generating images has never been as accesible as now!</figcaption></figure><h3>📑 Step 1: Set Up Your Project Folder</h3><ol><li>Open <strong>Finder</strong> and go to your <strong>Home</strong> directory.</li><li>Create a new folder named ai-projects.</li><li>This will be the base folder for all your <strong>Diffusion</strong> models, ComfyUI, and related tools.</li></ol><h3>💻 Step 2: Open Terminal and Learn Basic Commands</h3><p>Open <strong>Terminal</strong> from <strong>Applications → Utilities</strong>. Familiarize yourself with these useful commands:</p><pre>pwd         # Prints your current directory<br>ls          # Lists contents of the current directory<br>cd          # Changes directory</pre><p>To navigate to your ai-projects folder:</p><pre>cd ~/ai-projects</pre><p>Or simply drag the ai-projects folder into the terminal after typing cd .</p><h3>🍺 Step 3: Install Homebrew</h3><p>Homebrew is a package manager that simplifies software installations.</p><ol><li>Visit <a href="https://brew.sh/">brew.sh</a> and copy the installation command.</li><li>Paste it into Terminal and press <strong>Enter</strong>.</li><li>Enter your password when prompted.</li><li>If you see a “command not found” error, add Homebrew to your PATH:</li></ol><pre>echo &#39;eval &quot;$(/opt/homebrew/bin/brew shellenv)&quot;&#39; &gt;&gt; ~/.zprofile <br>eval &quot;$(/opt/homebrew/bin/brew shellenv)&quot;</pre><p>5. Test it with:</p><pre>brew doctor</pre><h3>🐍 Step 4: Install Python 3.11</h3><p>Check your current Python version:</p><pre>python3 --version</pre><p>If it’s not <strong>3.11</strong>, install it via Homebrew:</p><pre>brew install python@3.11</pre><h3>📦 Step 5: Install pip and setuptools</h3><p>Even though Python 3.11 often comes with pip, you can upgrade/install them to be sure:</p><pre>pip3 install --upgrade pip setuptools</pre><h3>🔥 Step 6: Install PyTorch (Nightly Build for Mac)</h3><p>To take advantage of Metal GPU support on Mac:</p><ol><li>Visit <a href="https://pytorch.org/get-started/locally/">PyTorch Get Started</a>.</li><li>Select: <strong>Mac → Pip → Python → Preview (Nightly)</strong></li><li>Copy the command and paste it into your terminal to install PyTorch.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MZ9XrK285pOGYyJ5Q4OsqA.png" /><figcaption>Options to be selected for the correct Pytorch version installation.</figcaption></figure><h3>🌐 Step 7: Clone the ComfyUI Repository</h3><ol><li>Navigate to your ai-projects directory (if not already there):</li></ol><pre>cd ~/ai-projects</pre><p>2. Clone ComfyUI from GitHub:</p><pre>git clone https://github.com/comfyanonymous/ComfyUI.git</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OD31lnD4rPVS3uDV8Jyymw.png" /><figcaption>ComfyUI’s GitHub repository page.</figcaption></figure><p>3. Enter the directory:</p><pre>cd ComfyUI</pre><h3>📋 Step 8: Install ComfyUI Dependencies</h3><p>Run the following to install the necessary Python libraries:</p><pre>pip3 install -r requirements.txt</pre><h3>🧠 Step 9: Launch ComfyUI</h3><p>Run the main Python script using Python 3.11:</p><pre>python3.11 main.py</pre><p>You’ll get a local URL like http://127.0.0.1:8188 . Open it in your browser to start using ComfyUI!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AqIqdkgX8cIO_C6v7an6NQ.png" /><figcaption>Opening screen of ComfyUI the first time you set it up.</figcaption></figure><h3>🔁 Reopening ComfyUI Later</h3><p>Whenever you restart your Mac or close the terminal:</p><pre>cd ~/ai-projects/ComfyUI<br>python3.11 main.py</pre><h3>🧰 Troubleshooting Tips</h3><ul><li>Confirm you’re using the correct Python version.</li><li>Revisit the official <a href="https://github.com/comfyanonymous/ComfyUI">ComfyUI GitHub repo</a> for updates.</li></ul><h3>🙌 Wrapping Up</h3><p>That’s it! You now have <strong>ComfyUI</strong> up and running on your Apple Silicon Mac! From here, you can start building your own workflows, experimenting with models, and generating stunning AI artwork or animations.</p><p>If you found this guide helpful, feel free to <strong>share it</strong> and connect with me over on <strong>X</strong>:<br> <a href="https://www.x.com/fortranibal"><strong>@fortranibal</strong></a> 🧠✨</p><blockquote><em>I’m also building </em><a href="https://www.tryreflektion.com/"><strong><em>Reflektion</em></strong></a><em>, an AI therapy assistant designed for meaningful conversations, emotional continuity, and deep personalization. Affordable, accessible, and always there for you.</em></blockquote><blockquote><strong><em>Find your path with Reflektion.</em></strong></blockquote><p>Happy creating! 🎨🤖</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a01f810a0ca3" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>