<?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 Matthias Grabner on Medium]]></title>
        <description><![CDATA[Stories by Matthias Grabner on Medium]]></description>
        <link>https://medium.com/@mattgrabner?source=rss-319d420ccf99------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*T0yjTtWKhTunKrpC6lFXiA@2x.jpeg</url>
            <title>Stories by Matthias Grabner on Medium</title>
            <link>https://medium.com/@mattgrabner?source=rss-319d420ccf99------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 10:25:20 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mattgrabner/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[#Dall-E2 prompt results for kids]]></title>
            <link>https://medium.com/@mattgrabner/dall-e2-prompt-results-for-kids-4f3c7a11ea55?source=rss-319d420ccf99------2</link>
            <guid isPermaLink="false">https://medium.com/p/4f3c7a11ea55</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[dalle-2]]></category>
            <dc:creator><![CDATA[Matthias Grabner]]></dc:creator>
            <pubDate>Tue, 27 Sep 2022 09:42:14 GMT</pubDate>
            <atom:updated>2022-09-27T09:44:32.122Z</atom:updated>
            <content:encoded><![CDATA[<p>Found out that when our little one is afraid or just unwilling to do or try certain things, she can be convinced by good examples of characters shown in children&#39;s books. If you are a parent you might know this all too well.</p><p>For those situations when you just can’t find a “best practice” in a book, OpenAI’s #Dall-E2 actually offers a great alternative — and thus I discovered yet another use case: #Dall-E2 prompts for kids!</p><p>The following prompts all end with “, illustration from children’s book”:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Nrqd-IMb_KlmRBVPQvzLyg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O6ZEneMxkronS6d4VaHhZA.png" /><figcaption>“Baby tiger making sand ice cream” and “Cute baby octopus riding a bicycle”</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jSwS3vFEVSj68N2Tqmm1Kw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y3JyG8ay9dLwKJvq1daY4g.png" /><figcaption>“Baby tiger riding a bicycle wearing a pink helmet” and “cute badger eating his snack sitting on a park bench”</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mmMS-Mf0hdvvVWyVZQMP9g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RWgVmUxalBgViBBDuJbQeA.png" /><figcaption>“cute bear running barefoot on the grass” and “cute bear swimming in the pool”</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iQ7yStfrHJMG4Dm3YVDMmw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*u4Mkp8GIz8aECoYGc_YgcQ.png" /><figcaption>“Cute baby bear with pink helmet riding a bike, kid’s painting” and “Mama badger and baby badger in the snow-covered forest”</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4f3c7a11ea55" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Stable-Diffusion on MB Pro Silicon M1 — Landscapes]]></title>
            <link>https://medium.com/@mattgrabner/stable-diffusion-on-mb-pro-silicon-m1-landscapes-459075eaa868?source=rss-319d420ccf99------2</link>
            <guid isPermaLink="false">https://medium.com/p/459075eaa868</guid>
            <category><![CDATA[stable-diffusion]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Matthias Grabner]]></dc:creator>
            <pubDate>Tue, 27 Sep 2022 09:30:14 GMT</pubDate>
            <atom:updated>2022-09-27T09:30:14.563Z</atom:updated>
            <content:encoded><![CDATA[<h3>Stable-Diffusion on MB Pro Silicon M1 — Landscapes</h3><blockquote>Prompt: “A beautiful alpine landscape under gorgeous skies, almost like in a dream”</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*oFgeM0SgeYVIEtELJ1T9aA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*wS7TDP1iH8Vc5DjBgC6IHQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*SJiEITiIg02KH2811RUpLA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*TwpNb3kbEWHnjSQt7tWkyA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*naBQgczCN0D88yzls3FAqA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*lCMoIuGZJzRpjEDOC7ZaTg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*RItAzF5fKPFnksrecXJirw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*JqWJzThM3UE8pZ6FVPMCLA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*KowLByJ4XdERnHHi7vRHVw.png" /></figure><blockquote>Prompt: “A beautiful landscape on mars under gorgeous martian skies, almost like in a dream”</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*ZkWC4OTz0Xgdr1qkCskEVw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*kH6aYTU6EcmclKO6hCX1nA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*OBbXzx6qfLGojrrTRNPa_Q.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*nXsr2sB5Nd93ff7szhVq0A.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*TdoQ8rX1wxpTC_rKsoIrSA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*pMLlCR-WJf3SLA9k_W-_HA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*TMwqoIO22ufXfZ_XfG06cw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*N45UG-F3SHydNGY195mNhQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*S4Wad8BcsNuXaHzLtZ8MqQ.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=459075eaa868" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Stable Diffusion on Macbook ProM1]]></title>
            <link>https://medium.com/@mattgrabner/stable-diffusion-on-macbook-prom1-949f7fe80f46?source=rss-319d420ccf99------2</link>
            <guid isPermaLink="false">https://medium.com/p/949f7fe80f46</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[stable-diffusion]]></category>
            <dc:creator><![CDATA[Matthias Grabner]]></dc:creator>
            <pubDate>Tue, 27 Sep 2022 08:53:35 GMT</pubDate>
            <atom:updated>2022-09-27T08:53:35.684Z</atom:updated>
            <content:encoded><![CDATA[<p>Stable Diffusion is a machine learning-based, text-to-image model to generate digital images from natural language text input. The underlying approach was developed at LMU Munich and then extended by a collaboration of StabilityAI, LMU, and Runway with support from EleutherAI and LAION. Unlike OpenAI’s Dall-E, Stable Diffusion is open source, which makes it even more revolutionary.</p><p>After its initial release on August 22, 2022, there are already hundreds of forks, versions, and implementations of Stable-Diffusion out there. Tinkerers are refining the models, training their own models, create adaptions for special use cases, even create plugins for software like Photoshop, Blender etc.</p><p>The great thing about stable-diffusion is that while it needs a lot of computing power you can also run it “at home”, here the current system requirements:</p><p><strong>System</strong></p><ul><li>An NVIDIA-based graphics card with 4 GB or more VRAM memory.</li><li>An Apple computer with an M1 chip.</li></ul><p><strong>Memory</strong></p><ul><li>At least 12 GB Main Memory RAM.</li></ul><p><strong>Disk</strong></p><ul><li>At least 6 GB of free disk space for the machine learning model, Python, and all its dependencies.</li></ul><p>If you are interested in trying it yourself, I suggest trying out this version of of CompVis/stable-diffusion, which features an interactive command-line script that combines text2img and img2img functionality in a “dream bot” style interface, a WebGUI, and multiple features and other enhancements: <a href="https://github.com/invoke-ai/InvokeAI">https://github.com/invoke-ai/InvokeAI</a></p><p>To make it run on my Macbook Pro M1 I referred to <a href="https://www.reddit.com/r/StableDiffusion/comments/x3yf9i/stable_diffusion_and_m1_chips_chapter_2/">this guide </a>posted on Reddit by Any-Winter-4079, it worked for me using this step-by-step method (beware, things are evolving quickly check the Git and comments on the Reddit thread for changes):</p><p>1. Open Terminal</p><p>2. Clone lstein’s repo, typing the command <em>git clone </em><a href="https://github.com/lstein/stable-diffusion.git"><em>https://github.com/lstein/stable-diffusion.git</em></a><em> </em>in your Terminal and clicking Enter. If you want to clone it in a specific folder, cd into it beforehand</p><p>3. Get into the project directory with <em>cd stable-diffusion</em>.</p><p>4. Create the conda environment with the command <em>CONDA_SUBDIR=osx-arm64 conda env create -f environment-mac.yaml</em>. If you get an Error because you already have an existing ldm environment, you can either update it, or you can open the environment-mac.yaml file that is inside your project directory in a text or code editor and change the first line from name: ldm to name: ldm-lstein or whatever new name you choose. Then in Terminal <em>CONDA_SUBDIR=osx-arm64 conda env create -f environment-mac.yaml.</em> This way you will preserve your original environment ldm and create a new one to test this new repo.</p><p>5. Activate the environment with the command<em> conda activate ldm</em> (or conda activate ldm-lstein or whatever environment name you chose in Step 4).</p><p>6. Place your sd-v1–4.ckpt weights in models/ldm/stable-diffusion-v1, where stable-diffusion-v1 is a new folder you create. Rename sd-v1–4.ckpt to model.ckpt. You can get these weights downloading sd-v1–4.ckpt from <a href="https://huggingface.co/CompVis/stable-diffusion-v-1-4-original">https://huggingface.co/CompVis/stable-diffusion-v-1-4-original</a> (note you will probably need to create an account and agree to Terms&amp;Conds)</p><p>7. Back in your Terminal, cd .. to get out of your project directory. Then, to add GFPGAN, use the command <em>git clone </em><a href="https://github.com/TencentARC/GFPGAN.git"><em>https://github.com/TencentARC/GFPGAN.git</em></a><em> </em>This should create a GFPGAN folder that is a sibling of your project folder (e.g. stable-diffusion).</p><p>8. Download GFPGANv1.3.pth from <a href="https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth">https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth</a></p><p>9. Put the file in experiments/pretrained_models folder, which is inside the GFPGAN folder (e.g. route GFPGAN/experiments/pretrained_models/)</p><p>10. Back in your Terminal, enter the GFPGAN folder with the command cd GFPGAN. We’ll be typing a few commands next.</p><p>11. <em>pip install basicsr</em></p><p>12. <em>pip install facexlib</em></p><p>13. <em>pip install -r requirements.txt</em></p><p>14. <em>python setup.py develop</em></p><p>15. <em>pip install realesrgan</em></p><p>16. After running these commands, you are ready to go. Type <em>cd ..</em> to get out of the GFPGAN folder, then cd stable-diffusion</p><p>17. <em>python3 scripts/preload_models.py</em></p><p>18. Finally, use the command <em>python3 ./scripts/dream.py</em> After initializing, you will see a dream &gt; prompt</p><p>19. Enter <em>Anubis riding a motorbike in Grand Theft Auto cover, palm trees, cover art by Stephen Bliss, artstation, high quality -m ddim -S 1805504473</em></p><p>20. You should be getting the following image if you are not using pytorch-nightly</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=949f7fe80f46" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AR / VR Landscape Austria]]></title>
            <link>https://medium.com/@mattgrabner/ar-vr-landscape-austria-f4baf2d21da5?source=rss-319d420ccf99------2</link>
            <guid isPermaLink="false">https://medium.com/p/f4baf2d21da5</guid>
            <category><![CDATA[austria]]></category>
            <category><![CDATA[industry-insights]]></category>
            <category><![CDATA[startup-ecosystem]]></category>
            <category><![CDATA[virtual-reality]]></category>
            <category><![CDATA[augmented-reality]]></category>
            <dc:creator><![CDATA[Matthias Grabner]]></dc:creator>
            <pubDate>Tue, 04 Feb 2020 10:08:53 GMT</pubDate>
            <atom:updated>2020-02-04T10:08:53.832Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>An overview on companies, platforms &amp; institutions leading the development of AR/VR technology in Austria.</strong></p><p>The <strong>AR / VR ecosystem in Austria</strong> is thriving, and covers the entire value chain: Hardware, software, development tools and SDKs. In our research, Clemens Wasner of Enlite AI and myself have identified 74 dedicated AR / VR companies. Dedicated AR /VR companies either develop their own hardware, software or SDKs, or significantly achieve added value to their products and services using technologies developed by third parties.</p><p><strong>You’re welcome to reuse the infographic below (first published </strong><a href="https://www.enlite.ai/works/ar-vr-landscape-austria"><strong>here</strong></a><strong>) as long as the content remains unmodified, in full and proper reference to the source is mentioned.</strong></p><figure><img alt="Austrian augmented reality and virtual reality landscape. An overview on companies, platforms &amp; institutions" src="https://cdn-images-1.medium.com/max/1024/1*8AC-1RLFTTsx0llG2RK87A.png" /></figure><p>We have clustered the landscapes in 2 larger groups: 1.) AR /VR Companies &amp; Startups containing 8 subgroups (Creative &amp; Media, Development &amp; Consulting, Solutions, Devices &amp; Hardware, Real Estate, Tools / Tech, Education &amp; Training and Industry); 2.) Enablers &amp; Extended Ecosystem (Universities &amp; Educational Institutions, Research and Communities)</p><h3><strong>Steady natural growth</strong></h3><p>While investment activity is steady, the majority of AR / VR companies on the landscape are self-funded and cash-flow positive. Compared to other landscapes AR / VR companies have been around for longer than their peers in Blockchain and AI.</p><h3>Industry, Health &amp; Smart City lead in R&amp;D</h3><p>Almost a third of the 30 AR / VR research projects we included in our landscape deal with Indutry specific applications. This resonates with the strong points of the Austrian economy, which lie in the machinery and automotive suppliyng industry. A quarter each of R&amp;D projects cover Health and Smart City solutions. In healthcare, AR / VR solutions mainly deal with solutions in the field of rehabilitative medicine, but also training &amp; education. Smart City solutions entail traffic information systems and city planning solutions.</p><p><em>Disclaimer: We are not involved with any of the presented companies and organizations listed in this landscape.</em></p><p><strong>You’re welcome to re-use the infographic below as long as the content remains unmodified, in full and proper reference to the source is mentioned.</strong></p><p><strong>Would you like your company or institution to be included in future updates? Simply send an email to </strong><a href="/service/mailto:office@enlite.ai"><strong>office@enlite.ai</strong></a><strong>, we’re happy to keep track of your request.</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f4baf2d21da5" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>