<?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/9ce1505034ab4e73a81274d8c2a04201&quot; frameborder=&quot;0&quot; width=&quot;1152&quot; height=&quot;864&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>864</height><width>1152</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>864</thumbnail_height><thumbnail_width>1152</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/9ce1505034ab4e73a81274d8c2a04201-ff692935e6892552.gif</thumbnail_url><duration>63.4227</duration><title>NNPort</title><description>In this video, I demonstrate a vibe debugging project aimed at identifying mismatches in neural network implementations across various platforms. I will upload a reference implementation, specifically the Python code, alongside a new C++ implementation. My goal is to compile and run the C++ code on OpenCL, then execute it on the target hardware connected via USB. This process will involve cross-compiling the OpenCL code, transferring it to the device, running it, and comparing the results with the reference implementation. I encourage viewers to follow along with the steps I outline to ensure accurate results.</description></oembed>