<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/6910578a2f5740daaf13807c78bbbe9a&quot; frameborder=&quot;0&quot; width=&quot;1850&quot; height=&quot;1387&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1387</height><width>1850</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1387</thumbnail_height><thumbnail_width>1850</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/6910578a2f5740daaf13807c78bbbe9a-00001.gif</thumbnail_url><duration>246.927</duration><title>Semantic Search for VASP Error Resolution</title><description>In this video, I explain how we have developed a model to assist users when they encounter errors with VASP, a widely used tool for DFT calculations. I discuss the limitations of the current VASP forum and documentation, and introduce our solution: a semantic search approach. We have created a database using data from the official VASP forum, and our model predicts the top 5 most similar issues and provides brief answers. This saves time and eliminates the need for manual searching. I also mention future extensions, such as adding configuration conversions and fine-tuning the model. Watch the video to learn more!</description></oembed>