<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Merry&apos;n&apos;Pippin</title><description>A lab notebook for small-budget robot learning, vision-language-action models, and embodied AI experiments — including the ones that fail.</description><link>https://merry-and-pippin.net/</link><language>en</language><item><title>Can a Robot Learn 81 Pick-and-Place Routes from Only 25?</title><link>https://merry-and-pippin.net/posts/can-a-robot-learn-81-pick-and-place-routes-from-only-25/</link><guid isPermaLink="true">https://merry-and-pippin.net/posts/can-a-robot-learn-81-pick-and-place-routes-from-only-25/</guid><description>Testing whether ACT, SmolVLA, and π0.5 can generalize to novel cube-to-cup routes after training on only 25 of 81 start-goal combinations.</description><pubDate>Mon, 14 Sep 2026 13:23:09 GMT</pubDate></item><item><title>Teaching Merry and Pippin a New Trick</title><link>https://merry-and-pippin.net/posts/teaching-merryn-and-pippin-a-new-trick/</link><guid isPermaLink="true">https://merry-and-pippin.net/posts/teaching-merryn-and-pippin-a-new-trick/</guid><description>Training an ACT imitation-learning policy for 20,000 steps and turning SO-101 demonstrations into autonomous movement with LeRobot.</description><pubDate>Wed, 26 Aug 2026 06:11:00 GMT</pubDate></item><item><title>The Commands Behind My LeRobot Workflow</title><link>https://merry-and-pippin.net/posts/the-commands-behind-my-lerobot-workflow/</link><guid isPermaLink="true">https://merry-and-pippin.net/posts/the-commands-behind-my-lerobot-workflow/</guid><description>The command-line workflow I use to collect LeRobot datasets, train ACT policies, and run ACT or SmolVLA inference on my two-camera SO-101 setup.</description><pubDate>Sun, 23 Aug 2026 13:02:00 GMT</pubDate></item><item><title>Giving My SO-101 Two Ways to See</title><link>https://merry-and-pippin.net/posts/giving-my-so-101-two-ways-to-see/</link><guid isPermaLink="true">https://merry-and-pippin.net/posts/giving-my-so-101-two-ways-to-see/</guid><description>How I equipped a LeRobot SO-101 with Logitech Brio 100 table cameras, a DFRobot FIT0701 wrist camera, and custom 3D-printed mounts.</description><pubDate>Tue, 18 Aug 2026 08:02:00 GMT</pubDate></item><item><title>Calibrating the LeRobot SO-101</title><link>https://merry-and-pippin.net/posts/calibrating-the-lerobot-so-101/</link><guid isPermaLink="true">https://merry-and-pippin.net/posts/calibrating-the-lerobot-so-101/</guid><description>Calibration was painless and straightforward. With the leader and follower assembled and connected, I followed the official Hugging Face LeRobot procedure and calibrated the two arms without any notable problems.</description><pubDate>Sun, 16 Aug 2026 12:20:00 GMT</pubDate></item><item><title>The Servo Was Innocent</title><link>https://merry-and-pippin.net/posts/the-servo-was-innocent/</link><guid isPermaLink="true">https://merry-and-pippin.net/posts/the-servo-was-innocent/</guid><description>The error appeared in software, so naturally I began by suspecting the most expensive piece of hardware.</description><pubDate>Sat, 15 Aug 2026 08:32:00 GMT</pubDate></item><item><title>Building My First LeRobot Arm</title><link>https://merry-and-pippin.net/posts/untitled-post/</link><guid isPermaLink="true">https://merry-and-pippin.net/posts/untitled-post/</guid><description>The structural parts were printed on a **Prusa MINI+** using PETG from [Prusament](https://prusament.com/pl/materials/prusament-petg/) and [ROSA3D](https://www.rosa3d.pl/filament/pet-g). PETG made sense for a working prototype: these were not display pieces but brackets, housings, links, and joints that would be handled, screwed together, taken apart, and—although I did not know it yet—occasionally printed again.</description><pubDate>Tue, 11 Aug 2026 18:05:00 GMT</pubDate></item></channel></rss>