--- title: NO GPU left behind: Unlocking Efficiency with Co-located... description: Hugging Face introduces co-located vLLM in TRL to solve GPU inefficiency in GRPO training. Previously, training and inference ran on separate GPUs, causing... canonical: https://daily.dev/posts/no-gpu-left-behind-unlocking-efficiency-with-co-located-vllm-in-trl-t6hkw5kdw twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: NO GPU left behind: Unlocking Efficiency with Co-located vLLM in TRL | daily.dev og:description: Hugging Face introduces co-located vLLM in TRL to solve GPU inefficiency in GRPO training. Previously, training and inference ran on separate GPUs, causing... og:url: https://daily.dev/posts/no-gpu-left-behind-unlocking-efficiency-with-co-located-vllm-in-trl-t6hkw5kdw og:image: https://api.daily.dev/og/posts/T6hkW5Kdw.png og:image:alt: NO GPU left behind: Unlocking Efficiency with Co-located vLLM in TRL 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. # NO GPU left behind: Unlocking Efficiency with Co-located vLLM in TRL **[Hugging Face](https://daily.dev/sources/huggingface)** · 14 min read · 0 upvotes · 0 comments ## Summary Hugging Face introduces co-located vLLM in TRL to solve GPU inefficiency in GRPO training. Previously, training and inference ran on separate GPUs, causing idle time and resource waste. The new approach runs both processes on the same GPUs, achieving up to 1.73× speedup while maintaining model quality. The solution includes vLLM's sleep() API for memory management, DeepSpeed ZeRO Stage 3 for large models, and seamless integration with distributed training tools. Successfully demonstrated on models up to 72B parameters with significant throughput improvements. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: --- [View this post on daily.dev](https://daily.dev/posts/no-gpu-left-behind-unlocking-efficiency-with-co-located-vllm-in-trl-t6hkw5kdw) ```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":"NO GPU left behind: Unlocking Efficiency with Co-located vLLM in TRL","url":"https://daily.dev/posts/no-gpu-left-behind-unlocking-efficiency-with-co-located-vllm-in-trl-t6hkw5kdw","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/no-gpu-left-behind-unlocking-efficiency-with-co-located-vllm-in-trl-t6hkw5kdw"},"datePublished":"2025-06-03T14:07:42.380Z","dateModified":"2025-06-03T14:08:06.163Z","description":"Hugging Face introduces co-located vLLM in TRL to solve GPU inefficiency in GRPO training. Previously, training and inference ran on separate GPUs, causing...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3adb585cc30566de533a80a9541b46d1?_a=AQAEulh","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3adb585cc30566de533a80a9541b46d1?_a=AQAEulh","isAccessibleForFree":true,"articleSection":"Hugging Face","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":"Hugging Face","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/f1f55c67d81a4330acf5b90b26b0c8e1","url":"https://daily.dev/sources/huggingface"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/no-gpu-left-behind-unlocking-efficiency-with-co-located-vllm-in-trl-t6hkw5kdw","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"","timeRequired":"PT14M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Hugging Face","item":"https://daily.dev/sources/huggingface"},{"@type":"ListItem","position":3,"name":"NO GPU left behind: Unlocking Efficiency with Co-located vLLM in TRL"}]} ```