{"version":1,"type":"story","url":"https://digestai.news/story/sensenova-sensenova-loop-9b-preview-hugging-face","json":"https://digestai.news/story/sensenova-sensenova-loop-9b-preview-hugging-face.json","markdown":"https://digestai.news/story/sensenova-sensenova-loop-9b-preview-hugging-face.md","slug":"sensenova-sensenova-loop-9b-preview-hugging-face","headline":"Sensenova/SenseNova-Loop-9B-Preview · Hugging Face","summary":"The model processes text and images, generating text for applications combining document understanding, coding, and external tool interactions. Its architecture includes a **looped decoder** with **two-stream manifold-constrained hyper-connections (mHC)**, where a 24-layer block is executed twice with shared weights, effectively doubling the computation path to 56 layers while reusing only 32 distinct layers. Training recipes use sequences up to **256,000 tokens**, incorporating text, images, code, and tool trajectories.\n\nThe model supports **MTP for accelerated inference**, with **Dspark and Dflash2** under development. Deployment requires a documented environment and configuration. The model is licensed under **Apache License 2.0** and released via Hugging Face. Downloads last month reached **12**.","keyPoints":["Uses a looped decoder with 24-layer blocks executed twice, doubling computation to 56 layers","Supports text, images, code, and tool trajectories with sequences up to 256K tokens"],"whyItMatters":"The model’s looped architecture and long-context support could improve workflows requiring iterative reasoning or multi-step tool interactions, appealing to developers and enterprises building agentic systems.","category":{"slug":"models","name":"Generative AI & Models","url":"https://digestai.news/category/models"},"entities":{"companies":["SenseNova"],"models":["SenseNova-Loop-9B-Preview"],"people":[]},"firstPublishedAt":"2026-10-30T12:34:31Z","updatedAt":"2026-10-30T12:34:31Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"huggingface.co","title":"sensenova/SenseNova-Loop-9B-Preview · Hugging Face","url":"https://huggingface.co/sensenova/SenseNova-Loop-9B-Preview","publishedAt":"2026-10-30T12:34:31Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[{"site":"Reddit","url":"https://www.reddit.com/r/LocalLLaMA/comments/1xje5oo/sensenovasensenovaloop9bpreview_hugging_face/","points":null}],"thread":null,"cite":{"text":"Digest AI, \"Sensenova/SenseNova-Loop-9B-Preview · Hugging Face\", 30 October 2026, https://digestai.news/story/sensenova-sensenova-loop-9b-preview-hugging-face","publisher":"Digest AI","title":"Sensenova/SenseNova-Loop-9B-Preview · Hugging Face","datePublished":"2026-10-30T12:34:31Z","url":"https://digestai.news/story/sensenova-sensenova-loop-9b-preview-hugging-face"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}