--- title: LeMario: Super Mario Bros trained on a JEPA Model description: A from-scratch implementation of a Joint-Embedding Predictive Architecture (JEPA) world model trained on Super Mario Bros gameplay footage. The model learns... canonical: https://daily.dev/posts/lemario-super-mario-bros-trained-on-a-jepa-model-vlkw5cnrf twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: LeMario: Super Mario Bros trained on a JEPA Model | daily.dev og:description: A from-scratch implementation of a Joint-Embedding Predictive Architecture (JEPA) world model trained on Super Mario Bros gameplay footage. The model learns... og:url: https://daily.dev/posts/lemario-super-mario-bros-trained-on-a-jepa-model-vlkw5cnrf og:image: https://api.daily.dev/og/posts/vLKw5CNrF.png og:image:alt: LeMario: Super Mario Bros trained on a JEPA Model og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # LeMario: Super Mario Bros trained on a JEPA Model **[Hacker News](https://daily.dev/sources/hn)** · 12 min read · 1 upvotes · 0 comments ## Summary A from-scratch implementation of a Joint-Embedding Predictive Architecture (JEPA) world model trained on Super Mario Bros gameplay footage. The model learns action-conditioned dynamics from pixels using a vision encoder, action encoder with AdaLN-Zero conditioning, and a causal transformer predictor. While the model successfully learned short-horizon dynamics and beat persistence baselines by 45.5% on five-step prediction, reward-free planning via the Cross-Entropy Method largely failed for distant goals. Key failure modes included the scrolling camera making visually similar but spatially distant locations indistinguishable in latent space, weak vertical position encoding, and a fundamental mismatch between predictive state and controllable state. The postmortem highlights that reproducing an architecture without matching its underlying dataset assumptions and environment properties (Push-T vs. Mario) leads to failure, and that validating central assumptions early is critical. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: --- Tags: [#reinforcement-learning](https://daily.dev/tags/reinforcement-learning) [View this post on daily.dev](https://daily.dev/posts/lemario-super-mario-bros-trained-on-a-jepa-model-vlkw5cnrf) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"LeMario: Super Mario Bros trained on a JEPA Model","url":"https://daily.dev/posts/lemario-super-mario-bros-trained-on-a-jepa-model-vlkw5cnrf","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/lemario-super-mario-bros-trained-on-a-jepa-model-vlkw5cnrf"},"datePublished":"2026-07-15T05:09:45.209Z","dateModified":"2026-07-18T04:21:24.086Z","description":"A from-scratch implementation of a Joint-Embedding Predictive Architecture (JEPA) world model trained on Super Mario Bros gameplay footage. The model learns...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/ca2510b35672e80918ef393e90a18133?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/ca2510b35672e80918ef393e90a18133?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Hacker News","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Hacker News","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/hn","url":"https://daily.dev/sources/hn"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/lemario-super-mario-bros-trained-on-a-jepa-model-vlkw5cnrf","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"reinforcement-learning","timeRequired":"PT12M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Hacker News","item":"https://daily.dev/sources/hn"},{"@type":"ListItem","position":3,"name":"LeMario: Super Mario Bros trained on a JEPA Model"}]} ```