[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-talmolab--sleap":3,"tool-talmolab--sleap":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":80,"owner_twitter":81,"owner_website":82,"owner_url":83,"languages":84,"stars":89,"forks":90,"last_commit_at":91,"license":92,"difficulty_score":10,"env_os":93,"env_gpu":94,"env_ram":95,"env_deps":96,"category_tags":103,"github_topics":104,"view_count":10,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":111,"updated_at":112,"faqs":113,"releases":142},991,"talmolab\u002Fsleap","sleap","A deep learning framework for multi-animal pose tracking.","SLEAP 是一个基于深度学习的开源框架，专注于多动物姿态跟踪。它能够精准追踪任意种类和数量的动物，并提供先进的标注与训练界面，帮助用户高效完成数据处理工作。无论是单只动物还是群体互动，SLEAP 都能通过自上而下或自下而上的训练策略实现高质量的姿态估计。\n\n在生物行为研究中，手动标注动物动作往往耗时且容易出错。SLEAP 解决了这一问题，不仅大幅减少人工干预，还能通过少量标注数据生成准确的预测结果。其快速训练和推理能力（单 GPU 训练仅需 15-60 分钟，推理速度可达每秒 600 帧以上）使其非常适合需要处理大规模视频数据的研究场景。\n\nSLEAP 主要面向生物学家、神经科学家以及从事动物行为分析的研究人员，同时也为开发者提供了灵活的 API 和独立的后端支持，便于集成到其他应用中。技术亮点包括可定制的神经网络架构、远程训练与推理支持，以及直观易用的图形界面。安装简单，跨平台兼容，适合从新手到高级用户的广泛人群使用。如果你正在寻找一种高效、精准的多动物姿态跟踪解决方案，SLEAP 将是一个值得尝试的选择。","[![CI](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Factions\u002Fworkflows\u002Fci.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Factions\u002Fworkflows\u002Fci.yml)\n[![Coverage](https:\u002F\u002Fcodecov.io\u002Fgh\u002Ftalmolab\u002Fsleap\u002Fbranch\u002Fdevelop\u002Fgraph\u002Fbadge.svg?token=oBmTlGIQRn)](https:\u002F\u002Fcodecov.io\u002Fgh\u002Ftalmolab\u002Fsleap)\n[![Documentation](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDocumentation-sleap.ai-lightgrey)](https:\u002F\u002Fdocs.sleap.ai)\n[![Downloads](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ftalmolab_sleap_readme_930e496e3e26.png)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fsleap)\n[![Stable version](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fv\u002Frelease\u002Ftalmolab\u002Fsleap?label=stable)](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002F)\n[![Latest version](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fv\u002Frelease\u002Ftalmolab\u002Fsleap?include_prereleases&label=latest)](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002F)\n\n# Social LEAP Estimates Animal Poses (SLEAP)\n\n![SLEAP Demo](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ftalmolab_sleap_readme_5508cff44dd5.gif)\n\n**SLEAP** is an open-source deep-learning based framework for multi-animal pose tracking [(Pereira et al., Nature Methods, 2022)](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41592-022-01426-1). It can be used to track any type or number of animals and includes an advanced labeling\u002Ftraining GUI for active learning and proofreading.\n\n## Features\n\n* Easy, one-line installation with support for all OSes\n* Purpose-built GUI and human-in-the-loop workflow for rapidly labeling large datasets\n* Single- and multi-animal pose estimation with *top-down* and *bottom-up* training strategies\n* Customizable neural network architectures that deliver *accurate predictions* with *very few* labels\n* Fast training: 15 to 60 mins on a single GPU for a typical dataset\n* Fast inference: up to 600+ FPS for batch, \u003C10ms latency for realtime\n* Support for remote training\u002Finference workflow (for using SLEAP without GPUs)\n* Flexible developer API for building integrated apps and customization\n* Two independent backends— [`sleap-nn`](https:\u002F\u002Fnn.sleap.ai) and [`sleap-io`](https:\u002F\u002Fio.sleap.ai) for training\u002Finference pipelines & handling SLEAP files respectively\n\n## Get some SLEAP\n\nSLEAP is installed as a Python package. We strongly recommend using [uv](https:\u002F\u002Fdocs.astral.sh\u002Fuv\u002F) to install SLEAP in its own environment.\n\nYou can find the latest version of SLEAP in the [Releases](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases) page.\n\n### Quick install\n\n> **Python 3.14 is not yet supported**\n>\n> SLEAP currently supports **Python 3.11, 3.12, and 3.13**.  \n> **Python 3.14 is not yet tested or supported.**  \n> By default, `uv` will use your system-installed Python.  \n> If you have Python 3.14 installed, you must specify the Python version (≤3.13) in the install command.  \n>\n> For example:\n>\n> ```bash\n> uv tool install --python 3.13 \"sleap[nn]\"  ...\n> ```\n> Replace `...` with the rest of your install command as needed.\n\n**`uv tool install` (any OS):**\n\nFirst, install [`uv`](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv) if you haven't already:\n\n```bash\n# macOS\u002FLinux\ncurl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\n\n# Windows\npowershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\n```\n\nThen install SLEAP:\n\n```bash\n# Windows\u002FLinux CUDA 12.8\nuv tool install \"sleap[nn]\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\n\n# macOS \u002F CPU-only\nuv tool install \"sleap[nn]\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\n```\n\nRun the SLEAP GUI after installation:\n\n```bash\nsleap\n```\n\nSee the docs for [full installation instructions](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Finstallation).\n\n## Learn to SLEAP\n\n- **Learn step-by-step:** [Tutorial](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Ftutorial\u002Foverview)\n- **Learn more advanced usage:** [Guides](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fguides\u002Fguides-overview\u002F) and [Notebooks](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fnotebooks\u002Fnotebooks-overview\u002F)\n- **Learn by watching:** [COSYNE 2024 Tutorial (Part 1)](https:\u002F\u002Fyoutu.be\u002FR5PRhkhAve0), [COSYNE 2024 Tutorial (Part 2)](https:\u002F\u002Fyoutu.be\u002FZ64v-vp-Jvo), [ABL:AOC 2023 Workshop](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=BfW-HgeDfMI), and [MIT CBMM Tutorial](https:\u002F\u002Fcbmm.mit.edu\u002Fvideo\u002Fdecoding-animal-behavior-through-pose-tracking)\n- **Learn by reading:** [Paper (Pereira et al., Nature Methods, 2022)](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41592-022-01426-1) and [Review on behavioral quantification (Pereira et al., Nature Neuroscience, 2020)](https:\u002F\u002Frdcu.be\u002FcaH3H)\n- **Learn from others:** [Discussions on Github](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fdiscussions)\n\n## References\n\nSLEAP is the successor to the single-animal pose estimation software [LEAP](https:\u002F\u002Fgithub.com\u002Ftalmo\u002Fleap) ([Pereira et al., Nature Methods, 2019](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41592-018-0234-5)).\n\nIf you use SLEAP in your research, please cite:\n\n> T.D. Pereira, N. Tabris, A. Matsliah, D. M. Turner, J. Li, S. Ravindranath, E. S. Papadoyannis, E. Normand, D. S. Deutsch, Z. Y. Wang, G. C. McKenzie-Smith, C. C. Mitelut, M. D. Castro, J. D'Uva, M. Kislin, D. H. Sanes, S. D. Kocher, S. S-H, A. L. Falkner, J. W. Shaevitz, and M. Murthy. [Sleap: A deep learning system for multi-animal pose tracking](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41592-022-01426-1). *Nature Methods*, 19(4), 2022\n\n**BibTeX:**\n\n```bibtex\n@ARTICLE{Pereira2022sleap,\n   title={SLEAP: A deep learning system for multi-animal pose tracking},\n   author={Pereira, Talmo D and \n      Tabris, Nathaniel and\n      Matsliah, Arie and\n      Turner, David M and\n      Li, Junyu and\n      Ravindranath, Shruthi and\n      Papadoyannis, Eleni S and\n      Normand, Edna and\n      Deutsch, David S and\n      Wang, Z. Yan and\n      McKenzie-Smith, Grace C and\n      Mitelut, Catalin C and\n      Castro, Marielisa Diez and\n      D'Uva, John and\n      Kislin, Mikhail and\n      Sanes, Dan H and\n      Kocher, Sarah D and\n      Samuel S-H and\n      Falkner, Annegret L and\n      Shaevitz, Joshua W and\n      Murthy, Mala},\n   journal={Nature Methods},\n   volume={19},\n   number={4},\n   year={2022},\n   publisher={Nature Publishing Group}\n   }\n}\n```\n\n\n**Technical issue with the software?**\n\n1. Check the [Help page](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fhelp).\n2. Ask the community via [discussions on Github](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fdiscussions).\n3. Search the [issues on GitHub](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fissues) or open a new one.\n\n**General inquiries?**\nReach out to [talmo@salk.edu](mailto:talmo@salk.edu).\n\n## Acknowledgments\n\nSLEAP was created in the [Murthy](https:\u002F\u002Fmurthylab.princeton.edu) and [Shaevitz](https:\u002F\u002Fshaevitzlab.princeton.edu) labs at the [Princeton Neuroscience Institute](https:\u002F\u002Fpni.princeton.edu) at Princeton University.\n\nSLEAP is currently being developed and maintained in the [Talmo Lab](https:\u002F\u002Ftalmolab.org) at the [Salk Institute for Biological Studies](https:\u002F\u002Fsalk.edu).\n\n### Maintainers\n\n* **Divya Murali** ([@gitttt-1234](https:\u002F\u002Fgithub.com\u002Fgitttt-1234)), Salk Institute for Biological Studies\n* **Amick Licup** ([@alicup29](https:\u002F\u002Fgithub.com\u002Falicup29)), Salk Institute for Biological Studies\n* **Elizabeth Berrigan** ([@eberrigan](https:\u002F\u002Fgithub.com\u002Feberrigan)), Salk Institute for Biological Studies\n* **Andrew Park** ([@7174Andy](https:\u002F\u002Fgithub.com\u002F7174Andy)), Salk Institute for Biological Studies\n* **Tom Han** ([@tom21100227](https:\u002F\u002Fgithub.com\u002Ftom21100227)), Salk Institute for Biological Studies\n* **Talmo Pereira** ([@talmo](https:\u002F\u002Fgithub.com\u002Ftalmo)), Salk Institute for Biological Studies\n\n### Contributors\nSLEAP would not be possible without major contributions from many folks, including:\n\n* **Liezl Maree**, Salk Institute for Biological Studies\n* **Arlo Sheridan**, Salk Institute for Biological Studies\n* **Arie Matsliah**, Princeton Neuroscience Institute, Princeton University\n* **Nat Tabris**, Princeton Neuroscience Institute, Princeton University\n* **David Turner**, Research Computing and Princeton Neuroscience Institute, Princeton University\n* **Joshua Shaevitz**, Physics and Lewis-Sigler Institute, Princeton University\n* **Mala Murthy**, Princeton Neuroscience Institute, Princeton University\n\n### Funding\n\nThis work was made possible through our funding sources, including:\n\n* NIH BRAIN Initiative R01 NS104899\n* Princeton Innovation Accelerator Fund\n* NIH BRAIN Initiative RF1 MH132653\n\n## License\n\n\u003C!-- SLEAP is released under a [Clear BSD License](https:\u002F\u002Fraw.githubusercontent.com\u002Ftalmolab\u002Fsleap\u002Fmain\u002FLICENSE) and is intended for research\u002Facademic use only. For commercial use, please contact: **Laurie Tzodikov (Assistant Director, Office of Technology Licensing), Princeton University, 609-258-7256**. -->\n\nSLEAP is released under a [BSD 3-Clause Clear License](LICENSE).\n\n## Links\n\n* [Documentation Homepage](https:\u002F\u002Fdocs.sleap.ai)\n* [Overview](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Foverview)\n* [Installation](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Finstallation)\n* [Tutorial](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Ftutorial\u002Foverview\u002F)\n* [Guides](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fguides\u002Fguides-overview\u002F)\n* [Notebooks](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fnotebooks\u002Fnotebooks-overview\u002F)\n* [Developer API](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fapi)\n* [Help](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fhelp)\n","[![CI](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Factions\u002Fworkflows\u002Fci.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Factions\u002Fworkflows\u002Fci.yml)\n[![Coverage](https:\u002F\u002Fcodecov.io\u002Fgh\u002Ftalmolab\u002Fsleap\u002Fbranch\u002Fdevelop\u002Fgraph\u002Fbadge.svg?token=oBmTlGIQRn)](https:\u002F\u002Fcodecov.io\u002Fgh\u002Ftalmolab\u002Fsleap)\n[![Documentation](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDocumentation-sleap.ai-lightgrey)](https:\u002F\u002Fdocs.sleap.ai)\n[![Downloads](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ftalmolab_sleap_readme_930e496e3e26.png)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fsleap)\n[![Stable version](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fv\u002Frelease\u002Ftalmolab\u002Fsleap?label=stable)](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002F)\n[![Latest version](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fv\u002Frelease\u002Ftalmolab\u002Fsleap?include_prereleases&label=latest)](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002F)\n\n# 社交型LEAP动物姿态估计系统（Social LEAP Estimates Animal Poses, SLEAP）\n\n![SLEAP演示](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ftalmolab_sleap_readme_5508cff44dd5.gif)\n\n**SLEAP** 是一个基于深度学习的开源多动物姿态追踪框架 [(Pereira et al., Nature Methods, 2022)](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41592-022-01426-1)。它可用于追踪任意类型或数量的动物，并包含专为活跃学习（active learning）和校对设计的高级标注\u002F训练图形用户界面（Graphical User Interface, GUI）。\n\n## 功能特性\n\n* 支持所有操作系统的简易单行安装\n* 专为快速标注大型数据集设计的图形用户界面（GUI）和人机协同工作流（human-in-the-loop workflow）\n* 支持单动物和多动物姿态估计，提供*自上而下（top-down）*和*自下而上（bottom-up）*训练策略\n* 可定制的神经网络架构，仅需*极少量*标注即可实现*高精度预测（accurate predictions）*\n* 快速训练：典型数据集在单GPU上训练耗时15至60分钟\n* 高速推理：批量处理可达600+帧\u002F秒（FPS），实时处理延迟低于10毫秒\n* 支持远程训练\u002F推理工作流（可在无GPU环境下使用SLEAP）\n* 灵活的开发者API，用于构建集成应用和自定义功能\n* 双独立后端——[`sleap-nn`](https:\u002F\u002Fnn.sleap.ai) 用于训练\u002F推理流程，[`sleap-io`](https:\u002F\u002Fio.sleap.ai) 用于处理SLEAP文件\n\n## 获取SLEAP\n\nSLEAP以Python包形式安装。强烈建议使用[uv](https:\u002F\u002Fdocs.astral.sh\u002Fuv\u002F)在独立环境中安装SLEAP。\n\n最新版本可在[发布版本](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases)页面获取。\n\n### 快速安装\n\n> **Python 3.14暂不支持**\n>\n> SLEAP当前支持**Python 3.11、3.12和3.13**。  \n> **Python 3.14尚未测试或支持。**  \n> 默认情况下，`uv`将使用系统安装的Python。  \n> 若已安装Python 3.14，必须在安装命令中指定Python版本（≤3.13）。  \n>\n> 示例：\n>\n> ```bash\n> uv tool install --python 3.13 \"sleap[nn]\"  ...\n> ```\n> 根据需要将`...`替换为安装命令的其余部分。\n\n**`uv tool install`（跨平台）：**\n\n若尚未安装[`uv`](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv)，请先安装：\n\n```bash\n# macOS\u002FLinux\ncurl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\n\n# Windows\npowershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\n```\n\n然后安装SLEAP：\n\n```bash\n# Windows\u002FLinux CUDA 12.8\nuv tool install \"sleap[nn]\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\n\n# macOS \u002F 仅CPU\nuv tool install \"sleap[nn]\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\n```\n\n安装后运行SLEAP GUI：\n\n```bash\nsleap\n```\n\n完整安装说明请参阅[文档](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Finstallation)。\n\n## 学习SLEAP\n\n- **逐步学习：** [教程](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Ftutorial\u002Foverview)\n- **进阶使用：** [指南](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fguides\u002Fguides-overview\u002F) 和 [示例笔记本](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fnotebooks\u002Fnotebooks-overview\u002F)\n- **视频学习：** [COSYNE 2024教程（第一部分）](https:\u002F\u002Fyoutu.be\u002FR5PRhkhAve0), [COSYNE 2024教程（第二部分）](https:\u002F\u002Fyoutu.be\u002FZ64v-vp-Jvo), [ABL:AOC 2023研讨会](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=BfW-HgeDfMI), 以及 [MIT CBMM教程](https:\u002F\u002Fcbmm.mit.edu\u002Fvideo\u002Fdecoding-animal-behavior-through-pose-tracking)\n- **文献学习：** [论文（Pereira et al., Nature Methods, 2022）](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41592-022-01426-1) 和 [行为量化综述（Pereira et al., Nature Neuroscience, 2020）](https:\u002F\u002Frdcu.be\u002FcaH3H)\n- **社区交流：** [GitHub讨论区](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fdiscussions)\n\n## 参考文献\n\nSLEAP是单动物姿态估计软件[LEAP](https:\u002F\u002Fgithub.com\u002Ftalmo\u002Fleap)的后续版本 [(Pereira et al., Nature Methods, 2019)](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41592-018-0234-5)。\n\n若在研究中使用SLEAP，请引用：\n\n> T.D. Pereira, N. Tabris, A. Matsliah, D. M. Turner, J. Li, S. Ravindranath, E. S. Papadoyannis, E. Normand, D. S. Deutsch, Z. Y. Wang, G. C. McKenzie-Smith, C. C. Mitelut, M. D. Castro, J. D'Uva, M. Kislin, D. H. Sanes, S. D. Kocher, S. S-H, A. L. Falkner, J. W. Shaevitz, and M. Murthy. [Sleap: A deep learning system for multi-animal pose tracking](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41592-022-01426-1). *Nature Methods*, 19(4), 2022\n\n**BibTeX:**\n\n```bibtex\n@ARTICLE{Pereira2022sleap,\n   title={SLEAP: A deep learning system for multi-animal pose tracking},\n   author={Pereira, Talmo D and \n      Tabris, Nathaniel and\n      Matsliah, Arie and\n      Turner, David M and\n      Li, Junyu and\n      Ravindranath, Shruthi and\n      Papadoyannis, Eleni S and\n      Normand, Edna and\n      Deutsch, David S and\n      Wang, Z. Yan and\n      McKenzie-Smith, Grace C and\n      Mitelut, Catalin C and\n      Castro, Marielisa Diez and\n      D'Uva, John and\n      Kislin, Mikhail and\n      Sanes, Dan H and\n      Kocher, Sarah D and\n      Samuel S-H and\n      Falkner, Annegret L and\n      Shaevitz, Joshua W and\n      Murthy, Mala},\n   journal={Nature Methods},\n   volume={19},\n   number={4},\n   year={2022},\n   publisher={Nature Publishing Group}\n   }\n}\n```\n\n\n**软件技术问题？**\n\n1. 查阅[帮助页面](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fhelp)。\n2. 通过[GitHub讨论区](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fdiscussions)向社区提问。\n3. 搜索[GitHub问题库](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fissues)或提交新问题。\n\n**常规咨询？**\n请联系 [talmo@salk.edu](mailto:talmo@salk.edu)。\n\n## 致谢\n\nSLEAP由普林斯顿大学普林斯顿神经科学研究所（[Princeton Neuroscience Institute](https:\u002F\u002Fpni.princeton.edu)）的[Murthy实验室](https:\u002F\u002Fmurthylab.princeton.edu)和[Shaevitz实验室](https:\u002F\u002Fshaevitzlab.princeton.edu)创建。\n\nSLEAP目前由索尔克生物研究所（[Salk Institute for Biological Studies](https:\u002F\u002Fsalk.edu)）的[Talmo实验室](https:\u002F\u002Ftalmolab.org)开发和维护。\n\n### 维护者\n\n* **Divya Murali** ([@gitttt-1234](https:\u002F\u002Fgithub.com\u002Fgitttt-1234)), Salk Institute for Biological Studies（索尔克生物研究所）\n* **Amick Licup** ([@alicup29](https:\u002F\u002Fgithub.com\u002Falicup29)), Salk Institute for Biological Studies\n* **Elizabeth Berrigan** ([@eberrigan](https:\u002F\u002Fgithub.com\u002Feberrigan)), Salk Institute for Biological Studies\n* **Andrew Park** ([@7174Andy](https:\u002F\u002Fgithub.com\u002F7174Andy)), Salk Institute for Biological Studies\n* **Tom Han** ([@tom21100227](https:\u002F\u002Fgithub.com\u002Ftom21100227)), Salk Institute for Biological Studies\n* **Talmo Pereira** ([@talmo](https:\u002F\u002Fgithub.com\u002Ftalmo)), Salk Institute for Biological Studies\n\n### 贡献者\nSLEAP 的开发离不开许多人的重大贡献，包括：\n\n* **Liezl Maree**, Salk Institute for Biological Studies\n* **Arlo Sheridan**, Salk Institute for Biological Studies\n* **Arie Matsliah**, Princeton Neuroscience Institute（普林斯顿神经科学研究所）, Princeton University（普林斯顿大学）\n* **Nat Tabris**, Princeton Neuroscience Institute, Princeton University\n* **David Turner**, Research Computing（研究计算）和 Princeton Neuroscience Institute, Princeton University\n* **Joshua Shaevitz**, Physics（物理学）和 Lewis-Sigler Institute（刘易斯-西格勒研究所）, Princeton University\n* **Mala Murthy**, Princeton Neuroscience Institute, Princeton University\n\n### 资金支持\n\n本工作得益于以下资金来源的支持：\n\n* NIH BRAIN Initiative（美国国立卫生研究院脑计划） R01 NS104899\n* Princeton Innovation Accelerator Fund（普林斯顿创新加速器基金）\n* NIH BRAIN Initiative（美国国立卫生研究院脑计划） RF1 MH132653\n\n## 许可证\n\n\u003C!-- SLEAP is released under a [Clear BSD License](https:\u002F\u002Fraw.githubusercontent.com\u002Ftalmolab\u002Fsleap\u002Fmain\u002FLICENSE) and is intended for research\u002Facademic use only. For commercial use, please contact: **Laurie Tzodikov (Assistant Director, Office of Technology Licensing), Princeton University, 609-258-7256**. -->\n\nSLEAP 在 [BSD 3-Clause Clear License（BSD 3条款清晰许可证）](LICENSE) 下发布。\n\n## 链接\n\n* [文档主页](https:\u002F\u002Fdocs.sleap.ai)\n* [概述](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Foverview)\n* [安装](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Finstallation)\n* [教程](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Ftutorial\u002Foverview\u002F)\n* [指南](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fguides\u002Fguides-overview\u002F)\n* [笔记本](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fnotebooks\u002Fnotebooks-overview\u002F)\n* [开发者API](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fapi)\n* [帮助](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fhelp)","# SLEAP 快速上手指南\n\n## 环境准备\n- **系统要求**：支持 Windows、macOS 和 Linux 所有主流操作系统\n- **Python 版本**：必须使用 Python 3.11、3.12 或 3.13（**不支持 Python 3.14**）\n- **前置依赖**：推荐通过 `uv` 工具创建独立环境；GPU 加速需 CUDA 12.8 支持（无 GPU 可用 CPU 模式）\n\n## 安装步骤\n1. **安装 uv 工具**（如未安装）：\n   ```bash\n   # macOS\u002FLinux\n   curl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\n\n   # Windows\n   powershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\n   ```\n\n2. **安装 SLEAP**：\n   - **GPU 版本（Windows\u002FLinux，CUDA 12.8）**：\n     ```bash\n     uv tool install \"sleap[nn]\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\n     ```\n   - **CPU 版本（macOS 或无 GPU）**：\n     ```bash\n     uv tool install \"sleap[nn]\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\n     ```\n   > **重要提示**：若系统存在 Python 3.14，需在命令中指定兼容版本（例如添加 `--python 3.13` 参数）\n\n3. **启动应用**：\n   ```bash\n   sleap\n   ```\n\n## 基本使用\n安装完成后，直接执行 `sleap` 命令启动图形界面。首次使用时：\n1. 导入示例视频（通过 `File > Open`）\n2. 使用标注工具框选动物关键点\n3. 点击 `Train` 按钮启动模型训练\n4. 训练完成后自动进行姿态跟踪并生成结果文件\n\n最简工作流：导入视频 → 标注 5-10 帧 → 训练 → 查看跟踪结果。详细操作可参考官方 [入门教程](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Ftutorial\u002Foverview)。","一位行为神经科学家在实验室中录制了10只C57BL\u002F6小鼠在开放场测试中的30分钟高清视频，需精确跟踪每只小鼠的鼻子、耳朵、脊柱和尾巴尖端等15个关键点，以量化社交距离和焦虑相关行为。\n\n### 没有 sleap 时\n- 手动标注视频帧极其耗时，标注一段30分钟视频（18,000帧）需专业人员耗时25小时以上，且小鼠重叠时个体ID错误率高达35%，导致数据不可靠。\n- 传统跟踪工具在动物追逐或遮挡时频繁丢失目标，数据链断裂严重，需额外8-10小时人工修复轨迹，影响行为分析的连续性。\n- 训练自定义模型需至少2,000帧高质量标注数据，收集过程耗时数周，训练占用高端GPU长达72小时，资源成本高昂。\n- 实时跟踪延迟超过100ms，无法在实验中即时触发刺激（如灯光变化），错失关键行为观察窗口。\n- 模型参数调整复杂，非技术背景研究人员难以优化，常因精度不足需反复重试，拖慢研究进度。\n\n### 使用 sleap 后\n- sleap的智能GUI通过主动学习自动建议关键点位置，标注时间缩短至3小时内，ID混淆率降至5%以下，确保数据高可靠性。\n- 其深度学习模型结合bottom-up策略有效处理遮挡场景，跟踪连续性提升85%，人工校正时间减少90%，行为数据更完整。\n- 仅需500帧标注数据即可训练高精度模型，训练在单GPU上1小时内完成，支持远程云服务加速，大幅降低资源门槛。\n- 推理速度达600+ FPS，延迟低于10ms，实现实时行为监测，助力实验中动态调整刺激参数。\n- 预置模型和简单API让研究人员快速部署定制方案，无需编程经验，模型精度显著提升且稳定。\n\nsleap将多动物姿态跟踪从繁琐低效的瓶颈转变为高效精准的研究引擎，使行为神经科学实验效率提升5倍以上。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Ftalmolab_sleap_5508cff4.gif","talmolab","Talmo Lab at the Salk Institute","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Ftalmolab_37c36aea.png","Talmo Pereira's research group at the Salk Institute for Biological Studies.",null,"talmo@salk.edu","talmop","talmolab.org","https:\u002F\u002Fgithub.com\u002Ftalmolab",[85],{"name":86,"color":87,"percentage":88},"Python","#3572A5",100,568,125,"2026-04-04T16:27:37","BSD-3-Clause-Clear","Linux, macOS, Windows","需要 NVIDIA GPU（CUDA 12.8），显存大小未说明；GPU 非必需，支持 CPU 模式","未说明",{"notes":97,"python":98,"dependencies":99},"必须使用 Python 3.11-3.13 版本（3.14 不支持）；推荐通过 uv 工具安装；macOS 仅支持 CPU 模式；GPU 非必需，但使用 NVIDIA GPU 时需 CUDA 12.8；安装时需指定 PyTorch 索引源","3.11, 3.12, 3.13",[100,101,102],"torch","sleap-nn","sleap-io",[13],[105,106,107,67,108,109,110],"pose-estimation","deep-learning","behavior-analysis","leap","animal-pose-estimation","animal-tracking","2026-03-27T02:49:30.150509","2026-04-06T05:36:32.516637",[114,119,124,128,133,137],{"id":115,"question_zh":116,"answer_zh":117,"source_url":118},4403,"使用修正帧训练新模型时出现错误怎么办？","该问题已在 SLEAP v1.1.2 版本中修复。请升级到最新版本。如果问题仍然存在，特别是对于 bottom-up 模型，请另开新 issue 报告。","https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fissues\u002F517",{"id":120,"question_zh":121,"answer_zh":122,"source_url":123},4404,"SLEAP 是否支持 RTX 3080 显卡？","是的，SLEAP v1.2.0 及以上版本支持 RTX 3080。请按照官方安装页面的说明进行设置：https:\u002F\u002Fsleap.ai\u002Finstallation.html","https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fissues\u002F454",{"id":125,"question_zh":126,"answer_zh":127,"source_url":123},4405,"安装后 GPU 无法识别，需要手动更新路径怎么办？","对于某些系统，可能需要添加 activate 脚本来更新 CUDA 路径。维护者计划在 conda 包中切换回 TensorFlow 2.6 以解决此问题。建议尝试安装 TensorFlow 2.6 版本的 SLEAP，例如使用命令 `conda create -y -n sleap -c sleap -c sleap\u002Flabel\u002Fdev -c nvidia -c conda-forge sleap=1.2.0a5`。",{"id":129,"question_zh":130,"answer_zh":131,"source_url":132},4406,"合并两个 SLEAP 项目时出现骨架冲突错误怎么办？","错误信息 'ValueError: Labels.skeleton can only be used when there is only a single skeleton saved in the labels' 表明项目中有多个骨架定义。确保在合并前两个项目使用完全相同的骨架。如果之前删除了节点，需重新定义骨架并保存为新 JSON 文件。","https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fissues\u002F713",{"id":134,"question_zh":135,"answer_zh":136,"source_url":132},4407,"打开包含多个视频的 SLP 文件时出错怎么办？","运行命令 `ulimit -n 2000` 在激活 SLEAP conda 环境后、打开 GUI 前，以增加系统文件描述符限制，解决打开多视频 SLP 文件的问题。",{"id":138,"question_zh":139,"answer_zh":140,"source_url":141},4408,"在 Google Colab 中训练模型时 GPU 不工作怎么办？","确保使用最新 SLEAP 版本，并在 Colab 中正确配置 GPU 运行时。如果问题持续，尝试重启运行时或切换 GPU 类型（如 Tesla T4 或 P100）。维护者确认此问题已解决，请参考最新文档。","https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fissues\u002F2479",[143,148,153,158,163,168,173,178,183,188,193,198,203,208,213,218,223,228,233,238],{"id":144,"version":145,"summary_zh":146,"released_at":147},103809,"v1.6.2","# SLEAP v1.6.2\r\n\r\nSLEAP v1.6.2 is a patch release with bug fixes for GUI performance, data integrity, training configuration, and updated dependencies.\r\n\r\n**Quick install\u002Fupgrade:**\r\n```bash\r\nuv tool install --python 3.13 \"sleap[nn]==1.6.2\" --torch-backend auto\r\n```\r\n\r\nSee the [v1.6.0 release notes](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0) for full details on the latest major release.\r\n\r\n---\r\n\r\n## Bug Fixes\r\n\r\n### Fix GUI freeze when adding instances on suggested frames (#2632)\r\n\r\nFixed a performance regression that caused the GUI to freeze for ~10 minutes when adding instances on videos with large suggestion sets (~100k suggestions). The status bar update had O(n×m) complexity which has been reduced to O(n+m).\r\n\r\n### Fix skeleton node removal not updating instance point data (#2621)\r\n\r\nFixed a critical bug where deleting nodes from a skeleton via the GUI did not update instance point arrays. This caused file corruption where saved files couldn't be reopened due to shape mismatches. Files now correctly update point data when nodes are removed.\r\n\r\n### Improve training dialog UI with smart field visibility (#2619)\r\n\r\nTraining dialog improvements:\r\n- Added toggle visibility for early stopping and OHKM parameter fields\r\n- Hide OHKM fields for centroid models where they don't apply\r\n- Fixed incorrect label \"Sigma for Edges\" → \"Sigma for Identity\" in multi-class bottom-up pipeline options\r\n\r\n### Fix Hydra override parsing error in exported train-script.sh (#2612)\r\n\r\nFixed `OverrideParseException` errors when running exported training scripts on SLURM clusters. The `ckpt_dir` and `run_name` values are now properly quoted to handle special characters.\r\n\r\n### Fix anchor part sync for top-down-id pipeline (#2610)\r\n\r\nFixed anchor_part selection not syncing correctly for the `top-down-id` pipeline in the training configuration dialog.\r\n\r\n---\r\n\r\n## Dependency Updates\r\n\r\n### sleap-io 0.6.4 → 0.6.5\r\n\r\n- **ROI and Segmentation Mask support (experimental)**: New `ROI` class for vector geometry and `SegmentationMask` class for raster masks\r\n- **COCO Detection & Segmentation I\u002FO (experimental)**: Read\u002Fwrite bounding box, polygon, and RLE mask annotations\r\n- **Ultralytics Detection & Segmentation I\u002FO (experimental)**: Extended YOLO format support\r\n- **NumPy 2.x compatibility**: Fixed serialization errors when saving `.slp` files\r\n\r\nSee the [sleap-io v0.6.5 release notes](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap-io\u002Freleases\u002Ftag\u002Fv0.6.5) for full details.\r\n\r\n### sleap-nn 0.1.0 → 0.1.2\r\n\r\n- **6.7x faster bottom-up inference** on NVIDIA A40 GPUs\r\n- **TUI Config Generator**: Interactive wizard for generating training configs on remote systems\r\n- **Post-processing filters**: New `--filter_min_visible_nodes`, `--filter_min_mean_node_score` options\r\n- **Multi-GPU fixes**: Fixed DDP collective mismatch crashes and NCCL deadlocks\r\n- **Tracking fixes**: Fixed track stealing bug and spurious track creation\r\n\r\nSee the [sleap-nn v0.1.1](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap-nn\u002Freleases\u002Ftag\u002Fv0.1.1) and [v0.1.2](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap-nn\u002Freleases\u002Ftag\u002Fv0.1.2) release notes for full details.\r\n\r\n---\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.6.1...v1.6.2\r\n","2026-03-07T00:30:45",{"id":149,"version":150,"summary_zh":151,"released_at":152},103810,"v1.6.1","# SLEAP v1.6.1\n\nSLEAP v1.6.1 is a patch release with bug fixes for Linux Qt compatibility, training configuration saving, and a new `--video-backend` CLI option.\n\n**Quick install\u002Fupgrade:**\n```bash\nuv tool install --python 3.13 \"sleap[nn]==1.6.1\" --torch-backend auto\n```\n\nSee the [v1.6.0 release notes](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0) for full details on the latest major release.\n\n---\n\n## Bug Fixes\n\n### Fix Linux Qt library conflicts (#2604)\n\nOn some Linux distributions (Debian 12, Fedora 43, and others with system Qt 6 packages), SLEAP could crash on launch with `ImportError: undefined symbol` errors due to conflicts between system Qt libraries and PySide6's bundled Qt. SLEAP now ensures PySide6's bundled Qt libraries take precedence on Linux by setting `LD_LIBRARY_PATH` and `QT_PLUGIN_PATH` before launching.\n\n### Fix training config save dialog bugs (#2603)\n\nFixed two bugs in the Training Configuration dialog reported in #2602:\n\n- **Run names were ignored when saving configs**: User-entered run names were overwritten with auto-generated timestamps when using \"Save configuration files...\" or \"Export training job package...\". Run names are now preserved correctly.\n- **YAML file picker showed no files**: The \"Select training config file...\" dropdown's file browser failed to show `.yaml`\u002F`.yml` files due to an incorrect file filter separator. Fixed to use Qt's expected format.\n\n---\n\n## New Features\n\n### `--video-backend` CLI option (#2604)\n\nA new `--video-backend` flag allows selecting the video decoding backend when launching SLEAP:\n\n```bash\nsleap --video-backend FFMPEG\n```\n\nThis is useful for working around h264 codec errors that can occur with OpenCV's default backend on some Linux systems. The setting persists across sessions via user preferences.\n\n---\n\n## Other Changes\n\n- Added `workflow_dispatch` trigger to the build workflow for manual CI re-runs\n- Added display\u002FGUI and video codec troubleshooting documentation\n\n---\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.6.0...v1.6.1\n","2026-02-10T01:31:32",{"id":154,"version":155,"summary_zh":156,"released_at":157},103811,"v1.6.0","# What's New in SLEAP 1.6\r\n\r\nSLEAP 1.6 is a major update with new backbone architectures, a redesigned training and inference experience, automated label quality control, ONNX\u002FTensorRT export for faster deployment, a unified CLI, and **MANY** bug fixes. This release spans **70+ PRs** since v1.5.2.\r\n\r\n**Quick start:**\r\n```bash\r\nuv tool install --python 3.13 \"sleap[nn]==1.6.0\" --torch-backend auto\r\n```\r\nSee below for more detailed installation instructions.\r\n\r\n\r\n## Major Changes\r\n\r\n### New Backbone Architectures in Training Dialog (#2579)\r\n\r\nSLEAP 1.6 brings ConvNeXt and Swin Transformer backbone support to the GUI training dialog, alongside UNet:\r\n\r\n- **ConvNeXt** -- Modern convolutional architecture in tiny\u002Fsmall\u002Fbase\u002Flarge variants (28M-198M parameters) with optional ImageNet pretrained weights for faster convergence.\r\n- **Swin Transformer (SwinT)** -- Transformer-based backbone in tiny\u002Fsmall\u002Fbase variants (28M-88M parameters) with optional ImageNet pretrained weights.\r\n\r\nSelect these from the training dialog's new backbone selector dropdown. See the [sleap-nn model documentation](https:\u002F\u002Fnn.sleap.ai\u002Flatest\u002Fmodels\u002F#backbone-architectures) for details.\r\n\r\n### Unified `sleap` CLI\r\n\r\nSLEAP now has a single `sleap` command as the primary entry point. Running `sleap` launches the GUI, and subcommands provide access to all tools:\r\n\r\n```bash\r\nsleap              # Launch the GUI\r\nsleap doctor       # System diagnostics and troubleshooting\r\nsleap export-model # Export models to ONNX\u002FTensorRT\r\n```\r\n\r\nAll sleap-io and sleap-nn CLI commands are now integrated as `sleap` subcommands (#2524, #2541, #2559, #2587, #2595, #2597):\r\n\r\n| Command | Description |\r\n|---------|-------------|\r\n| `sleap doctor` | System diagnostics and troubleshooting |\r\n| `sleap show` | Display labels file summary |\r\n| `sleap convert` | Convert between label formats |\r\n| `sleap split` | Split labels into train\u002Fval\u002Ftest |\r\n| `sleap unsplit` | Recombine split labels |\r\n| `sleap merge` | Merge multiple labels files |\r\n| `sleap render` | Render pose videos |\r\n| `sleap fix` | Fix\u002Frepair labels files |\r\n| `sleap embed` \u002F `sleap unembed` | Manage embedded video data |\r\n| `sleap trim` | Trim labels to subset |\r\n| `sleap reencode` | Re-encode embedded videos |\r\n| `sleap transform` | Coordinate-aware video transformations |\r\n| `sleap filenames` | List video filenames in labels |\r\n| `sleap train` | Train models (from sleap-nn) |\r\n| `sleap predict` | Run inference (from sleap-nn) |\r\n| `sleap export-model` | Export models to ONNX\u002FTensorRT |\r\n\r\n### ONNX & TensorRT Model Export\r\n\r\nExport trained models to optimized formats for **3-6x faster inference** (#2573, #2594, #2595, #2597):\r\n\r\n```bash\r\nsleap export-model model.ckpt -o model.onnx --format onnx\r\nsleap export-model model.ckpt -o model.engine --format tensorrt\r\n```\r\n\r\nRun inference on exported models:\r\n\r\n```bash\r\nsleap predict model.onnx video.mp4 -o predictions.slp\r\n```\r\n\r\nBenchmark results (NVIDIA RTX A6000, batch size 8):\r\n\r\n| Model Type | PyTorch | TensorRT FP16 | Speedup |\r\n|------------|---------|---------------|---------|\r\n| Single Instance | 3,111 FPS | 11,039 FPS | **3.5x** |\r\n| Centroid | 453 FPS | 1,829 FPS | **4.0x** |\r\n| Top-Down | 94 FPS | 525 FPS | **5.6x** |\r\n| Bottom-Up | 113 FPS | 524 FPS | **4.6x** |\r\n\r\nTo install export dependencies, reinstall with the appropriate extra: `\"sleap[nn,export]==1.6.0\"` (ONNX), `\"sleap[nn,export-gpu]==1.6.0\"` (ONNX + GPU), or `\"sleap[nn,tensorrt]==1.6.0\"` (TensorRT). See the [sleap-nn Export Guide](https:\u002F\u002Fnn.sleap.ai\u002Flatest\u002Fexport\u002F) for full benchmarks and details.\r\n\r\n### Label Quality Control (#2547)\r\n\r\nNew `sleap.qc` module with GMM-based anomaly detection to automatically identify annotation errors. Accessible via **Analyze > Label QC...** in the GUI.\r\n\r\n- Detects 10+ error types: isolated misses, jitter, visibility errors, scale issues, left-right swaps, gross misses, missing instances, and duplicates\r\n- Dockable GUI widget with score histograms and sensitivity controls\r\n- Keyboard navigation (Space\u002FShift+Space) to quickly review flagged instances\r\n- Export to CSV or add flagged instances to Suggestions for batch review\r\n\r\n### Redesigned Training & Inference Dialogs (#2506, #2509, #2519, #2556, #2557, #2579)\r\n\r\nThe training and inference dialogs have been completely redesigned with native Qt for a faster, more polished experience:\r\n\r\n- **55x faster config loading** via rapidyaml and lazy loading\r\n- **Smaller dialog** that fits on 1280x720 screens\r\n- **Augmentation controls** simplified with on\u002Foff checkboxes and rotation presets (default: full ±180°)\r\n- **Device and worker settings** default from user preferences\r\n- **Random Seed field** for reproducible train\u002Fvalidation splits\r\n- **Evaluation metrics** can be computed at configurable epoch intervals (mOKS, mAP, mAR, PCK, distance metrics logged to WandB)\r\n- **Prediction handling modes**: Choose Keep, Replace, or Clear all predictions during inference\r\n- **\"Random sample (current video)\"** inference targe","2026-02-06T18:30:03",{"id":159,"version":160,"summary_zh":161,"released_at":162},103812,"v1.6.0a3","# SLEAP v1.6.0a3\r\n\r\n## About the v1.6 Pre-release Series\r\n\r\nThis is a **pre-release** for SLEAP v1.6.0. It contains many new features and improvements, but is not yet considered stable. For production use, see [v1.5.2](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.5.2).\r\n\r\nWe are releasing a series of pre-releases that incrementally build towards the stable v1.6.0 release. Each pre-release adds new features and bug fixes:\r\n\r\n| Version                                                                 | Summary                                                                                                                     |\r\n| ----------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- |\r\n| [**v1.6.0a0**](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0a0) | Unified `sleap` CLI, redesigned training dialog (55x faster loading), bug fixes for adding instances from predictions       |\r\n| [**v1.6.0a1**](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0a1) | Label QC for automated error detection, 8 new CLI commands from sleap-io, video rendering with live preview                 |\r\n| [**v1.6.0a2**](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0a2) | Revamped installation docs, epoch-end evaluation metrics, content-based video matching, bug fix for export training package |\r\n| **v1.6.0a3** *(current)*                                                | ConvNeXt\u002FSwinT backbones, ONNX\u002FTensorRT export, real-time inference progress, macOS bug fixes |\r\n\r\n> **Note**: Starting with SLEAP v1.5+, all deep learning functionality is powered by the PyTorch-based `sleap-nn` backend. TensorFlow models (with `UNet` backbones) from earlier versions are still supported for inference. Refer to the [Migrating to 1.5+](https:\u002F\u002Fdocs.sleap.ai\u002Fv1.6.0a3\u002Fguides\u002Fmigrating-to-sleap-1-5\u002F) docs for more details!\r\n\r\n---\r\n\r\n## How to Install\r\n\r\n**Step 1**: Install [`uv`](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv) (skip if already installed)\r\n\r\n```bash\r\n# Windows\r\npowershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\r\n\r\n# macOS\u002FLinux\r\ncurl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\r\n```\r\n\r\n**Step 2**: Install SLEAP v1.6.0a3\r\n\r\n```bash\r\nuv tool install --python 3.13 \"sleap[nn]==1.6.0a3\" --with \"sleap-io==0.6.3\" --with \"sleap-nn==0.1.0a4\" --prerelease allow --torch-backend auto\r\n```\r\n\r\nThat's it! SLEAP is now available system-wide. The `--torch-backend auto` flag automatically detects your GPU (NVIDIA, AMD, Intel, or CPU). Be sure to do a `uv self update` if you get an error about this flag.\r\n\r\n**Step 3**: Verify installation\r\n```bash\r\nsleap doctor\r\n```\r\n\r\n### Upgrading from v1.6.0a2?\r\n\r\n```bash\r\nuv tool upgrade sleap --upgrade-package sleap-io --upgrade-package sleap-nn\r\n```\r\n\r\nOr for a clean reinstall:\r\n\r\n```bash\r\nuv tool install --reinstall --python 3.13 \"sleap[nn]==1.6.0a3\" --with \"sleap-io==0.6.3\" --with \"sleap-nn==0.1.0a4\" --prerelease allow --torch-backend auto\r\n```\r\n\r\n### Rollback to stable\r\n\r\nIf you encounter issues, rollback to the latest stable release:\r\n\r\n```bash\r\nuv tool install --python 3.13 \"sleap[nn]==1.5.2\" --torch-backend auto\r\n```\r\n\r\n### Version compatibility\r\n\r\n| SLEAP   | sleap-io | sleap-nn |\r\n| ------- | -------- | -------- |\r\n| 1.6.0a3 | 0.6.3    | 0.1.0a4  |\r\n| 1.6.0a2 | 0.6.2    | 0.1.0a2  |\r\n| 1.6.0a1 | 0.6.1    | 0.1.0a1  |\r\n\r\n---\r\n\r\n## What's New in v1.6.0a3\r\n\r\n### New Backbone Architectures ([#2579](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2579))\r\n\r\nTrain models using modern transformer-based and ConvNeXt architectures as alternatives to UNet:\r\n\r\n- **ConvNeXt**: Select from tiny\u002Fsmall\u002Fbase\u002Flarge variants (28M-198M parameters) with optional ImageNet pretrained weights for faster convergence\r\n- **Swin Transformer (SwinT)**: Select from tiny\u002Fsmall\u002Fbase variants (28M-88M parameters) with optional ImageNet pretrained weights\r\n\r\nAccess these from the training dialog's new backbone selector dropdown.\r\n\r\n### ONNX\u002FTensorRT Model Export ([#2573](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2573))\r\n\r\nExport trained models to optimized formats for **3-6x faster inference**:\r\n\r\n```bash\r\nsleap-nn-export model.ckpt -o model.onnx --format onnx\r\nsleap-nn-export model.ckpt -o model.engine --format tensorrt\r\n```\r\n\r\nRun inference on exported models:\r\n\r\n```bash\r\nsleap-nn-predict model.onnx video.mp4 -o predictions.slp\r\n```\r\n\r\n**Benchmark results** (NVIDIA RTX A6000, batch size 8):\r\n\r\n| Model Type | PyTorch | TensorRT FP16 | Speedup |\r\n|------------|---------|---------------|---------|\r\n| Single Instance | 3,111 FPS | 11,039 FPS | **3.5x** |\r\n| Centroid | 453 FPS | 1,829 FPS | **4.0x** |\r\n| Top-Down | 94 FPS | 525 FPS | **5.6x** |\r\n| Bottom-Up | 113 FPS | 524 FPS | **4.6x** |\r\n\r\nSee the [sleap-nn Export Guide](https:\u002F\u002Fnn.sleap.ai\u002Fprerelease\u002Fexport\u002F) for full benchmarks and usage details.\r\n\r\n### Real-Time Inference Progress ([#2575](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2575))\r\n\r\nThe inference dialo","2026-01-21T00:50:01",{"id":164,"version":165,"summary_zh":166,"released_at":167},103813,"v1.6.0a2","# SLEAP v1.6.0a2\n\n## About the v1.6 Pre-release Series\n\nThis is a **pre-release** for SLEAP v1.6.0. It contains many new features and improvements, but is not yet considered stable. For production use, see [v1.5.2](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.5.2).\n\nWe are releasing a series of pre-releases that incrementally build towards the stable v1.6.0 release. Each pre-release adds new features and bug fixes:\n\n| Version                                                                 | Summary                                                                                                                     |\n| ----------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- |\n| [**v1.6.0a0**](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0a0) | Unified `sleap` CLI, redesigned training dialog (55x faster loading), bug fixes for adding instances from predictions       |\n| [**v1.6.0a1**](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0a1) | Label QC for automated error detection, 8 new CLI commands from sleap-io, video rendering with live preview                 |\n| **v1.6.0a2** *(current)*                                                | Revamped installation docs, epoch-end evaluation metrics, content-based video matching, bug fix for export training package |\n\n> **Note**: Starting with SLEAP v1.5+, all deep learning functionality is powered by the PyTorch-based `sleap-nn` backend. TensorFlow models (with `UNet` backbones) from earlier versions are still supported for inference. Refer to the [Migrating to 1.5+](https:\u002F\u002Fdocs.sleap.ai\u002Fv1.6.0a2\u002Fguides\u002Fmigrating-to-sleap-1-5\u002F) docs for more details!\n\n---\n\n## How to Install\n\n**Step 1**: Install [`uv`](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv) (skip if already installed)\n\n```bash\n# Windows\npowershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\n\n# macOS\u002FLinux\ncurl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\n```\n\n**Step 2**: Install SLEAP v1.6.0a2\n\n```bash\nuv tool install --python 3.13 \"sleap[nn]==1.6.0a2\" --with \"sleap-io==0.6.2\" --with \"sleap-nn==0.1.0a2\" --prerelease allow --torch-backend auto\n```\n\nThat's it! SLEAP is now available system-wide. The `--torch-backend auto` flag automatically detects your GPU (NVIDIA, AMD, Intel, or CPU). Be sure to do a `uv self update` if you get an error about this flag.\n\n**Step 3**: Verify installation\n```bash\nsleap doctor\n```\n\n### Upgrading from v1.6.0a1?\n\n```bash\nuv tool upgrade sleap --upgrade-package sleap-io --upgrade-package sleap-nn\n```\n\nOr for a clean reinstall:\n\n```bash\nuv tool install --reinstall --python 3.13 \"sleap[nn]==1.6.0a2\" --with \"sleap-io==0.6.2\" --with \"sleap-nn==0.1.0a2\" --prerelease allow --torch-backend auto\n```\n\n### Rollback to stable\n\nIf you encounter issues, rollback to the latest stable release:\n\n```bash\nuv tool install --python 3.13 \"sleap[nn]==1.5.2\" --torch-backend auto\n```\n\n### Version compatibility\n\n| SLEAP   | sleap-io | sleap-nn |\n| ------- | -------- | -------- |\n| 1.6.0a2 | 0.6.2    | 0.1.0a2  |\n| 1.6.0a1 | 0.6.1    | 0.1.0a1  |\n| 1.6.0aN | 0.6.x    | 0.1.0aN  |\n| 1.6.x   | 0.6.x    | 0.1.x    |\n| 1.5.x   | 0.5.x    | 0.0.x    |\n\n---\n\n## What's New in v1.6.0a2\n\n- **Revamped Installation Documentation:**\n    - Complete rewrite of installation docs with simplified workflow (#2567)\n    - Single universal install command for all platforms using `--torch-backend auto`\n    - Reduced from 8 installation paths to 2 (tool install + development setup)\n    - New `uvx sleap labels.slp` option for viewing data without permanent installation\n    - Streamlined upgrade flow with `uv tool upgrade sleap`\n\n- **Python 3.13 Default:**\n    - Python 3.13 is now the default recommended version (#2565)\n    - Python 3.12 remains supported\n\n- **Bug Fixes:**\n    - Fixed \"Export Training Job Package\" crash with `ConfigAttributeError: Missing key zmq` that occurred on v1.6.0a0\u002Fv1.6.0a1 (#2566, fixes #2562)\n\n- **[sleap-io v0.6.2](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap-io\u002Freleases\u002Ftag\u002Fv0.6.2):**\n    - **Content-Based Video Matching**: Videos are now automatically matched by pose annotations or pixel content, enabling reliable cross-platform merges even when file paths differ\n    - **New `Labels.match()` API**: Inspect matching results without merging — ideal for evaluation workflows\n    - **Video Color Mode Control**: New `Labels.set_video_color_mode()` method and `sio fix --video-color` CLI option\n    - Bug fixes for HDF5 dataset matching and provenance conflict handling\n\n- **[sleap-nn v0.1.0a2](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap-nn\u002Freleases\u002Ftag\u002Fv0.1.0a2):**\n    - **Epoch-End Evaluation Metrics**: Real-time mOKS, mAP, mAR, PCK, and distance metrics logged to WandB during training\n    - **Robust Video Matching**: Uses sleap-io's `Labels.match()` API for better cross-platform evaluation\n    - Bug fixes for embedded video handling and centroid model ground truth matching\n\n---\n\n### Full ","2026-01-16T21:25:08",{"id":169,"version":170,"summary_zh":171,"released_at":172},103814,"v1.6.0a1","# SLEAP v1.6.0a1\n\n## About the v1.6 Pre-release Series\n\nThis is a **pre-release** for SLEAP v1.6.0. It contains many new features and improvements, but is not yet considered stable. For production use, see [v1.5.2](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.5.2).\n\nWe are releasing a series of pre-releases that incrementally build towards the stable v1.6.0 release. Each pre-release adds new features and bug fixes:\n\n| Version | Summary |\n|---------|---------|\n| [**v1.6.0a0**](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0a0) | Unified `sleap` CLI, redesigned training dialog (55x faster loading), bug fixes for adding instances from predictions |\n| **v1.6.0a1** *(current)* | Label QC for automated error detection, 8 new CLI commands from sleap-io, video rendering with live preview |\n\n> **Note**: Starting with SLEAP v1.5+, all deep learning functionality is powered by the PyTorch-based `sleap-nn` backend. TensorFlow models (with `UNet` backbones) from earlier versions are still supported for inference. Refer to the [Migrating to 1.5+](https:\u002F\u002Fdocs.sleap.ai\u002Fv1.6.0a1\u002Fguides\u002Fmigrating-to-sleap-1-5\u002F) docs for more details!\n\n---\n\n## How to Install\n\n**Step 1**: Install [`uv`](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv) (skip if already installed)\n\n```bash\n# Windows\npowershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\n\n# macOS\u002FLinux\ncurl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\n```\n\n**Step 2**: Install SLEAP v1.6.0a1\n\n#### Windows\u002FLinux with NVIDIA GPU (CUDA 12.8)\n```bash\nuv tool install --reinstall --python 3.12 \"sleap[nn]==1.6.0a1\" --with \"sleap-io==0.6.1\" --with \"sleap-nn==0.1.0a1\" --prerelease allow --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\n```\n\n#### Windows\u002FLinux with NVIDIA GPU (CUDA 13.0)\n```bash\nuv tool install --reinstall --python 3.12 \"sleap[nn]==1.6.0a1\" --with \"sleap-io==0.6.1\" --with \"sleap-nn==0.1.0a1\" --prerelease allow --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu130 --index https:\u002F\u002Fpypi.org\u002Fsimple\n```\n\n#### Windows\u002FLinux without GPU (CPU only)\n```bash\nuv tool install --reinstall --python 3.12 \"sleap[nn]==1.6.0a1\" --with \"sleap-io==0.6.1\" --with \"sleap-nn==0.1.0a1\" --prerelease allow --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\n```\n\n#### macOS\n```bash\nuv tool install --reinstall --python 3.12 \"sleap[nn]==1.6.0a1\" --with \"sleap-io==0.6.1\" --with \"sleap-nn==0.1.0a1\" --prerelease allow\n```\n\n**Step 3**: Verify installation\n```bash\nsleap doctor\n```\n\n### Upgrading from v1.6.0a0?\n\nUse the same commands as above. The `--reinstall` flag will create a clean environment with the new dependencies.\n\n### Rollback to stable\n\nIf you encounter issues, rollback to the latest stable release:\n\n```bash\n# Windows\u002FLinux (CUDA 12.8)\nuv tool install --reinstall --python 3.12 \"sleap[nn]==1.5.2\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\n\n# Windows\u002FLinux (CPU only)\nuv tool install --reinstall --python 3.12 \"sleap[nn]==1.5.2\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\n\n# macOS\nuv tool install --reinstall --python 3.12 \"sleap[nn]==1.5.2\"\n```\n\n### Version compatibility\n\n| SLEAP | sleap-io | sleap-nn |\n|-------|----------|----------|\n| 1.6.0a1 | 0.6.1 | 0.1.0a1 |\n| 1.6.0aN | 0.6.x | 0.1.0aN |\n| 1.6.x | 0.6.x | 0.1.x |\n| 1.5.x | 0.5.x | 0.0.x |\n\n---\n\n## What's New in v1.6.0a1\n\n- **Label Quality Control (QC):**\n    - New `sleap.qc` module with GMM-based anomaly detection to automatically identify annotation errors (#2547)\n    - Detects 10+ error types: isolated misses, jitter, visibility errors, scale issues, left-right swaps, gross misses, missing instances, and duplicates\n    - Dockable GUI widget accessible via **Analyze > Label QC...** with score histograms and sensitivity controls\n    - Keyboard navigation (Space\u002FShift+Space) to quickly navigate flagged instances\n    - Export to CSV or add flagged instances to Suggestions for review\n\n- **Enhanced CLI:**\n    - 8 new CLI commands from sleap-io: `sleap merge`, `sleap unsplit`, `sleap fix`, `sleap embed`, `sleap unembed`, `sleap trim`, `sleap reencode`, `sleap transform` (#2559)\n    - See the [sleap-io CLI documentation](https:\u002F\u002Fio.sleap.ai\u002Fv0.6.1\u002Fcli\u002F) for detailed usage\n\n- **Video Rendering Overhaul:**\n    - Now powered by sleap-io's rendering engine — see [rendering documentation](https:\u002F\u002Fio.sleap.ai\u002Fv0.6.1\u002Frendering\u002F) for details (#2558)\n    - Live preview of rendered frames with all style options before exporting\n    - 12+ new color palettes and 5 marker shapes with options to color by track, instance, or node\n    - Alpha transparency support for overlays\n    - Non-blocking video export with progress bar and cancel support\n\n- **Training Dialog Improvements:**\n    - Form state now persists after clicking Cancel (#2557)\n    - Device and worker settings default from user preferences instead of being overwritten by profiles (#2557)\n    - Updated all baseline profiles to use full ±180° rotation augmentation (#2557)\n    - Added Random Seed ","2026-01-13T18:38:33",{"id":174,"version":175,"summary_zh":176,"released_at":177},103815,"v1.6.0a0","# SLEAP v1.6.0a0\n\n## About the v1.6 Pre-release Series\n\nThis is a **pre-release** for SLEAP v1.6.0. It contains many new features and improvements, but is not yet considered stable. For production use, see [v1.5.2](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.5.2).\n\nWe are releasing a series of pre-releases that incrementally build towards the stable v1.6.0 release. Each pre-release adds new features and bug fixes:\n\n| Version | Summary |\n|---------|---------|\n| **v1.6.0a0** *(current)* | Unified `sleap` CLI, redesigned training dialog (55x faster loading), bug fixes for adding instances from predictions |\n| [**v1.6.0a1**](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.6.0a1) | Label QC for automated error detection, 8 new CLI commands from sleap-io, video rendering with live preview |\n\n> **Note**: Starting with SLEAP v1.5+, all deep learning functionality is powered by the PyTorch-based `sleap-nn` backend. TensorFlow models (with `UNet` backbones) from earlier versions are still supported for inference. Refer to the [Migrating to 1.5+](https:\u002F\u002Fdocs.sleap.ai\u002Fv1.6.0a0\u002Fguides\u002Fmigrating-to-sleap-1-5\u002F) docs for more details!\n\n---\n\n## How to Install\n\n**Step 1**: Install [`uv`](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv) (skip if already installed)\n\n```bash\n# Windows\npowershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\n\n# macOS\u002FLinux\ncurl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\n```\n\n**Step 2**: Install SLEAP v1.6.0a0\n\n#### Windows\u002FLinux with NVIDIA GPU (CUDA 12.8)\n```bash\nuv tool install --force --python 3.12 \"sleap[nn]==1.6.0a0\" --with \"sleap-io==0.6.0\" --with \"sleap-nn==0.1.0a0\" --prerelease allow --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\n```\n\n#### Windows\u002FLinux with NVIDIA GPU (CUDA 13.0 - NEW!)\n```bash\nuv tool install --force --python 3.12 \"sleap[nn]==1.6.0a0\" --with \"sleap-io==0.6.0\" --with \"sleap-nn==0.1.0a0\" --prerelease allow --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu130 --index https:\u002F\u002Fpypi.org\u002Fsimple\n```\n\n#### Windows\u002FLinux without GPU (CPU only)\n```bash\nuv tool install --force --python 3.12 \"sleap[nn]==1.6.0a0\" --with \"sleap-io==0.6.0\" --with \"sleap-nn==0.1.0a0\" --prerelease allow --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\n```\n\n#### macOS\n```bash\nuv tool install --force --python 3.12 \"sleap[nn]==1.6.0a0\" --with \"sleap-io==0.6.0\" --with \"sleap-nn==0.1.0a0\" --prerelease allow\n```\n\n**Step 3**: Verify installation\n```bash\nsleap doctor\n```\n\n### Upgrading from v1.5.x?\n\nUse the same commands as above. The `--force` flag will replace your existing installation.\n\n### Rollback to stable\n\nIf you encounter issues, rollback to the latest stable release:\n\n```bash\n# Windows\u002FLinux (CUDA 12.8)\nuv tool install --force --python 3.12 \"sleap[nn]==1.5.2\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\n\n# Windows\u002FLinux (CPU only)\nuv tool install --force --python 3.12 \"sleap[nn]==1.5.2\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\n\n# macOS\nuv tool install --force --python 3.12 \"sleap[nn]==1.5.2\"\n```\n\n### Version compatibility\n\n| SLEAP | sleap-io | sleap-nn |\n|-------|----------|----------|\n| 1.6.0aN | 0.6.x | 0.1.0aN |\n| 1.6.x | 0.6.x | 0.1.x |\n| 1.5.x | 0.5.x | 0.0.x |\n\n---\n\n## What's New in v1.6.0a0\n\n- **Unified CLI:**\n    - New `sleap` command as the primary entry point - just run `sleap` to launch the GUI (#2524)\n    - Integrated sleap-io CLI commands: `sleap show`, `sleap convert`, `sleap split`, `sleap filenames`, `sleap render` (#2541)\n    - New `sleap doctor` command for system diagnostics and troubleshooting (#2524)\n\n- **Training GUI Overhaul:**\n    - **55x faster config loading** and much faster dialog startup (#2506, #2516)\n    - Completely redesigned training dialog with native Qt, unified frame targeting, and 356 new tests (#2519)\n    - New Frame Target Selector for flexible training\u002Finference frame selection (#2519)\n    - Prediction handling modes: Keep, Replace, or Clear all predictions during inference (#2519)\n    - Smaller dialog that fits on 1280x720 screens (#2509, #2519)\n    - Augmentation controls simplified with on\u002Foff checkboxes and rotation presets (#2509)\n    - WandB integration improvements with run URL display and auto-browser-open (#2525)\n\n- **Crop Size Visualization:**\n    - Visual crop size preview in training dialog for top-down models (#2483)\n    - New Instance Size Distribution widget for analyzing bounding box sizes across your dataset (#2528)\n    - Click-to-navigate from size distribution plot to specific instances (#2528)\n\n- **New Features:**\n    - \"Check for Updates\" dialog showing versions for sleap, sleap-io, and sleap-nn (#2499)\n    - \"Delete Predictions on User-Labeled Frames\" for cleaning up duplicate instances (#2505)\n    - Startup banner with version info and branding when launching `sleap-label` (#2517)\n    - Progress dialog with cancel support for package export (#2522)\n    - Support for loading legacy SLEAP metrics from v1.4.1 and earlier (#2","2026-01-11T20:02:16",{"id":179,"version":180,"summary_zh":181,"released_at":182},103816,"v1.5.2","## What's Changed\r\n\r\n### SLEAP v1.5.2 – Bug Fixes & Dependency Updates\r\n\r\nThis release includes important bug fixes for GUI rendering and Windows compatibility, dependency updates for improved stability, and further documentation improvements.\r\n\r\n> **Note**: Starting with SLEAP v1.5+, all deep learning functionality is powered by the PyTorch-based `sleap-nn` backend. TensorFlow models (with `UNet` backbones) from earlier versions are still supported for inference. Refer [Migrating to 1.5+](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fguides\u002Fmigrating-to-sleap-1-5\u002F) docs for more details!\r\n\r\n### How to install?\r\n\r\nYou can now install SLEAP quickly using [uv](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv)\r\n\r\n**Step 1**: Install [`uv`](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv) - an ultra-fast Python package manager\r\n\r\n```bash\r\n# Windows\r\npowershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\r\n\r\n# macOS\u002FLinux\r\ncurl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\r\n ```\r\n\r\n**Step 2**: Install `sleap`\r\n\r\n```bash\r\n# Windows\u002F Linux (CUDA)\r\nuv tool install --python 3.13 \"sleap[nn]==1.5.2\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\r\n\r\n# Windows\u002F Linux (CPU)\r\nuv tool install --python 3.13 \"sleap[nn]==1.5.2\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\r\n\r\n# macOS\r\nuv tool install --python 3.13 \"sleap[nn]==1.5.2\"\r\n\r\n ```\r\nCheck the full [installation guide](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Finstallation\u002F) for platform-specific instructions and advanced options.\r\n\r\nOnce you've installed SLEAP, run the below command from anywhere in your terminal\r\n```bash\r\nsleap-label\r\n```\r\nThe GUI should open up!\r\n\r\n---\r\n\r\n### Upgrading from v1.5.1?\r\n\r\nIf you already have SLEAP v1.5.1 installed, you can upgrade to v1.5.2 using the following commands based on your installation method:\r\n\r\n#### If installed with `uv tool install`:\r\n\r\nThe simplest upgrade command (preserves your original Python version and index URLs):\r\n```bash\r\nuv tool upgrade sleap\r\n```\r\n\r\nOr, if you want to ensure you're using Python 3.13 and refresh your installation:\r\n```bash\r\nuv tool uninstall sleap\r\n# Then reinstall with the commands from the installation section above\r\n```\r\n\r\n> **Note**: `uv tool upgrade` automatically preserves the index URLs (CUDA\u002FCPU) and Python version from your original installation. If you installed with `--index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128`, the upgrade will continue using the CUDA 12.8 index.\r\n\r\n#### If installed with `pip` in a conda environment:\r\n```bash\r\nconda activate sleap\r\npip install --upgrade \"sleap[nn]\"\r\n```\r\n\r\nFor platform-specific indexes (CUDA\u002FCPU), add the appropriate `--extra-index-url`:\r\n```bash\r\n# CUDA 12.8\r\npip install --upgrade \"sleap[nn]\" --extra-index-url https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index-url https:\u002F\u002Fpypi.org\u002Fsimple\r\n\r\n# CPU\r\npip install --upgrade \"sleap[nn]\" --extra-index-url https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index-url https:\u002F\u002Fpypi.org\u002Fsimple\r\n```\r\n\r\n#### If installed with `uv add` (project-based):\r\n```bash\r\n# Navigate to your project directory\r\nuv sync --upgrade\r\n```\r\n\r\n#### If installed from source:\r\n```bash\r\ncd sleap\r\ngit pull\r\nuv sync --upgrade\r\n```\r\n\r\nAfter upgrading, verify the installation:\r\n```bash\r\npython -c \"import sleap; sleap.versions()\"\r\n```\r\n\r\nYou should see `SLEAP: 1.5.2` in the output.\r\n\r\n---\r\n\r\n### Highlights\r\n\r\n- **Dependency updates:**\r\n    - Updated minimum `sleap-io` version to 0.5.7\r\n    - Updated minimum `sleap-nn` version to 0.0.4\r\n    - Removed `cattrs` dependency for simplified dependency management\r\n    - Added `--python 3.13` flag to installation commands to prevent Python 3.14 compatibility issues\r\n\r\n- **Bug fixes:**\r\n    - **Fixed color rendering in `sleap-render`**: Videos now display correct colors with proper BGR to RGB conversion (#2444)\r\n    - **Fixed Windows GUI crash**: Resolved Qt widget attribute error when loading .slp files on Windows (#2440)\r\n    - **Fixed instance coloring**: Multiple instances in older SLEAP projects now display with distinct colors instead of the same color (#2434)\r\n\r\n- **Documentation improvements:**\r\n    - Consolidated repetitive installation documentation (reduced by 55 lines while preserving all essential information)\r\n    - Improved `uv add` installation workflow instructions with Windows troubleshooting tips\r\n    - Clearer platform-specific installation guidance\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.5.1...v1.5.2\r\n","2025-10-30T03:57:03",{"id":184,"version":185,"summary_zh":186,"released_at":187},103817,"v1.5.1","## What's Changed\r\n\r\n### SLEAP v1.5.1 – Bug fixes & Documentation Improvements\r\n\r\nThis release focuses on a few bug fixes in the training pipeline, improving installation instructions, and updating documentation for a smoother user experience.\r\n\r\n> **Note**: Starting with SLEAP v1.5+, all deep learning functionality is powered by the PyTorch-based `sleap-nn` backend. TensorFlow models (with `UNet` backbones) from earlier versions are still supported for inference. Refer [Migrating to 1.5+](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Fguides\u002Fmigrating-to-sleap-1-5\u002F) docs for more details!\r\n\r\n### How to install?\r\n\r\nYou can now install SLEAP quickly using [uv](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv)\r\n\r\n**Step 1**: Install [`uv`](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv) - an ultra-fast Python package manager\r\n\r\n```bash\r\n# Windows\r\npowershell -c \"irm https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.ps1 | iex\"\r\n\r\n# macOS\u002FLinux\r\ncurl -LsSf https:\u002F\u002Fastral.sh\u002Fuv\u002Finstall.sh | sh\r\n ```\r\n\r\n**Step 2**: Install `sleap`\r\n\r\n```bash\r\n# Windows\u002F Linux (CUDA)\r\nuv tool install \"sleap[nn]\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcu128 --index https:\u002F\u002Fpypi.org\u002Fsimple\r\n\r\n# Windows\u002F Linux (CPU)\r\nuv tool install \"sleap[nn]\" --index https:\u002F\u002Fdownload.pytorch.org\u002Fwhl\u002Fcpu --index https:\u002F\u002Fpypi.org\u002Fsimple\r\n\r\n# macOS\r\nuv tool install \"sleap[nn]\"\r\n\r\n ```\r\nCheck the full [installation guide](https:\u002F\u002Fdocs.sleap.ai\u002Flatest\u002Finstallation\u002F) for platform-specific instructions and advanced options.\r\n\r\nOnce you've installed SLEAP, run the below command from anywhere in your terminal\r\n```bash\r\nsleap-label\r\n```\r\nThe GUI should open up!\r\n\r\n---\r\n\r\n### Highlights\r\n\r\n- **Improved installation:**\r\n    - Platform-specific dependency groups for sleap installation with CUDA support.\r\n    - Fixed CUDA installation issues on Windows.\r\n    - Updated installation instructions and options for clarity.\r\n- **Documentation updates:**\r\n    - Fixed typos and broken links.\r\n    - Improved CLI docs with new options and guidance on legacy CLIs.\r\n    - Fixed MkDocs versioning and improved doc site structure.\r\n- **Error handling:** sleap-nn import errors are now handled gracefully with clear user guidance.\r\n- **Bug fixes:** Minor fixes across CLI and docs to improve stability.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.5.0...v1.5.1","2025-10-02T19:11:46",{"id":189,"version":190,"summary_zh":191,"released_at":192},103818,"v1.5.0","# What's New in SLEAP 1.5\r\n\r\nSLEAP 1.5 represents a major milestone with significant architectural improvements, performance enhancements, and new installation methods. Here are the key changes:\r\n\r\n## Major Changes\r\n\r\n### Updated dependencies\r\nWe have now updated to support Python 3.12+ and support many new versions of the many libraries that SLEAP uses. This should make it much easier to install on modern platforms, support new architectures, and make development much easier.\r\n\r\n### UV-Based Installation\r\nSLEAP 1.5+ now uses [**uv**](https:\u002F\u002Fdocs.astral.sh\u002Fuv\u002F) for installation, making it much faster than previous methods. Get up and running in seconds with our streamlined installation process.\r\n\r\n### PyTorch Backend\r\nNeural network backend switched from TensorFlow to PyTorch via [`sleap-nn`](https:\u002F\u002Fnn.sleap.ai), providing:\r\n\r\n- **Much faster training and inference speeds**: Up to 2.5x faster training and inference times.\r\n- **Modern deep learning capabilities**: PyTorch with upcoming integrations with a whole slew of modern deep learning models and packages.\r\n- **Improved developer experience**: Check out the dedicated backend repo at https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap-nn\r\n- **Multi-GPU training**: Full support for using multiple GPUs for accelerated and larger scale training.\r\n- **Backwards compatibility**: You are able to use your existing trained SLEAP models from v1.4.1 for the UNet backend with no changes (see notes below).\r\n\r\n### Refreshed Documentation Websites\r\n- The new landing page is now live at: https:\u002F\u002Fsleap.ai\r\n- The new documentation is now live at: https:\u002F\u002Fdocs.sleap.ai\r\n- The old docs (v1.4.1) are will remain available at: https:\u002F\u002Flegacy.sleap.ai\r\n\r\n### Standalone Libraries\r\nSLEAP GUI is now supported by two new packages for modular workflows:\r\n\r\n#### [SLEAP-IO](https:\u002F\u002Fio.sleap.ai)\r\nI\u002FO backend for handling labels, processing `.slp` files, and data manipulation. Essential for any SLEAP workflow and can be used independently for data processing tasks.\r\n\r\n#### [SLEAP-NN](https:\u002F\u002Fnn.sleap.ai)\r\nPyTorch-based neural network backend for training and inference. Perfect for custom training pipelines, remote processing, and headless server deployments.\r\n\r\n## Torch Backend Changes\r\n\r\n### New Backbones\r\nSLEAP 1.5 introduces three powerful new backbone architectures (check [here](https:\u002F\u002Fnn.sleap.ai\u002Flatest\u002Fmodels\u002F#backbone-architectures) for more details):\r\n\r\n- **UNet** - Classic encoder-decoder architecture for precise pose estimation\r\n- **SwinT** - Swin Transformer for state-of-the-art performance\r\n- **ConvNeXt** - Modern convolutional architecture with improved efficiency\r\n\r\n### Legacy Support\r\nWe've maintained full backward compatibility:\r\n\r\n- **GUI Support**: SLEAP now uses a new \u003Cu>YAML-based\u003C\u002Fu> config file structure, but you can still upload and work with old SLEAP JSON files in the GUI. For details on converting legacy SLEAP 1.4 config\u002FJSON files to the new YAML format, see our [conversion guide](https:\u002F\u002Fnn.sleap.ai\u002Flatest\u002Fconfig\u002F#converting-legacy-sleap-14-configjson-to-sleap-nn-yaml).\r\n- **TensorFlow Model Inference**: Continue to support running inference on old TensorFlow models (UNet backbone only). Check [using legacy models](https:\u002F\u002Fnn.sleap.ai\u002Flatest\u002Finference\u002F#legacy-sleap-model-support) for more details.\r\n\r\n","2025-09-30T21:44:44",{"id":194,"version":195,"summary_zh":196,"released_at":197},103819,"v1.4.1","SLEAP 1.4.1 releases many new changes since the last big release 1.3.3. We hope users enjoy these long awaited new features and fixes!\r\n\r\n#### From 1.3.2+, to install SLEAP through pip use `pip install sleap[pypi]` to ensure all dependencies are gathered.\r\n\r\nAs a reminder:\r\n\r\n> The 1.3.1 dependency update requires [Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Findex.html) for faster dependency resolution. If you already have anaconda installed, then you _can_ set the solver to libmamba in the base environment:\r\n>```\r\n>conda update -n base conda\r\n>conda install -n base conda-libmamba-solver\r\n>conda config --set solver libmamba\r\n>```\r\n>Any subsequent `mamba` commands in the docs will need to be replaced with `conda` if you choose to use your existing Anaconda installation. \r\n>\r\n>Otherwise, follow the recommended installation instruction for [Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Findex.html).\r\n\r\n# Quick install\r\n**`mamba` (Windows\u002FLinux\u002FGPU)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c nvidia -c sleap\u002Flabel\u002Fdev -c sleap -c anaconda sleap=1.4.1\r\n```\r\n\r\n**`mamba` (Mac)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c anaconda -c sleap sleap=1.4.1\r\n```\r\n\r\n**`pip` (any OS except Apple Silicon)**:\r\n```\r\npip install sleap[pypi]==1.4.1\r\n```\r\n\r\n# Highlights\r\n* Add options to set background color when exporting video by @scott-yj-yang in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1328\r\n* Add resize\u002Fscroll to training GUI by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1565\r\n* Highlight instance box on hover by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2055\r\n* Enable touchpad pinch to zoom by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2058\r\n* Do not always color skeletons table black by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1952\r\n* Make status bar dependent on UI mode by @7174Andy in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2063\r\n* Graceful failing with seeking errors by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1712\r\n* Import DLC with uniquebodyparts, add Tracks by @getzze in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1562\r\n* Fix GUI crash on scroll by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1883\r\n\r\n# Full Changelog\r\n\r\n## Enhancements\r\n* Add options to set background color when exporting video by @scott-yj-yang in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1328\r\n* Increase range on batch size by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1513\r\n* Add resize\u002Fscroll to training GUI by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1565\r\n* support loading slp files with non-compound types and str in metadata by @lambdaloop in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1566\r\n* change inference pipeline option to tracking-only by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1666\r\n* Only propagate Transpose Tracks when propagate is checked by @vaibhavtrip29 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1748\r\n* Add batch size to GUI for inference by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1771\r\n* Add ZMQ support via GUI and CLI by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1780\r\n* Change menu name to match deleting predictions beyond max instance by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1790\r\n* Adding ragged metadata to `info.json` by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1765\r\n* Add option to export to CSV via sleap-convert and API by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1730\r\n* Add `normalized_instance_similarity` method  by @gitttt-1234 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1939\r\n* Update installation docs 1.4.1 by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1810\r\n* Option for Max Stride to be 128 by @MweinbergUmass in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1941\r\n* Allow csv and text file support on sleap track by @emdavis02 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1875\r\n* Added Three Different Cases for Adding a New Instance by @7174Andy in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1859\r\n* Generate suggestions using max point displacement threshold by @gqcpm in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1862\r\n* Add object keypoint similarity method by @getzze in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1003\r\n* Allowing inference on multiple videos via `sleap-track` by @emdavis02 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1784\r\n* Add `Keep visualizations` checkbox to training GUI by @hajin-park in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1824\r\n* Menu option to open preferences directory and update to util functions to pathlib by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1843\r\n* Add tracking score as seekbar header options by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2047\r\n* Don't mark complete on instance scaling by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2049\r\n* Add check for instances with track assigned before training ID models by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F2053\r\n* Add menu item for deleting instances beyond frame limit b","2024-12-19T17:57:58",{"id":199,"version":200,"summary_zh":201,"released_at":202},103820,"v1.3.4","SLEAP 1.3.4 has no changes to the SLEAP source code, but adds constraints to the `attrs` and `opencv` versions being pulled in.\r\n\r\n#### From 1.3.2+, to install SLEAP through pip use `pip install sleap[pypi]` to ensure all dependencies are gathered.\r\n\r\nAs a reminder:\r\n\r\n> The 1.3.1 dependency update requires [Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Findex.html) for faster dependency resolution. If you already have anaconda installed, then you _can_ set the solver to libmamba in the base environment:\r\n>```\r\n>conda update -n base conda\r\n>conda install -n base conda-libmamba-solver\r\n>conda config --set solver libmamba\r\n>```\r\n>Any subsequent `mamba` commands in the docs will need to be replaced with `conda` if you choose to use your existing Anaconda installation. \r\n>\r\n>Otherwise, follow the recommended installation instruction for [Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Findex.html).\r\n\r\n# Quick install\r\n**`mamba` (Windows\u002FLinux\u002FGPU)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c nvidia -c sleap -c anaconda sleap=1.3.4\r\n```\r\n\r\n**`mamba` (Mac)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c anaconda -c sleap sleap=1.3.4\r\n```\r\n\r\n**`pip` (any OS except Apple Silicon)**:\r\n```\r\npip install sleap[pypi]==1.3.4\r\n```\r\n\r\n# Full Changelog\r\n- Constrain attrs (mac) and opencv (linux) in 1.3.4 #1927","2024-09-03T20:49:20",{"id":204,"version":205,"summary_zh":206,"released_at":207},103821,"v1.4.1a2","SLEAP v1.4.1a2 is a pre-release. See [1.3.3](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.3.3) for the latest stable release. The crucial change here is _Fix zmq inference by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1800_  since inference was not working  in the pre-release v1.4.1a1 due to the addition of zmq port options for training in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1780 that were not being used for inference.\r\n\r\n# Quick install\r\n**`mamba` (Windows\u002FLinux\u002FGPU)**:\r\n```\r\nmamba create -y -n sleap_v1.4.1a2 -c conda-forge -c nvidia -c sleap\u002Flabel\u002Fdev -c anaconda sleap=1.4.1a2\r\n```\r\n\r\n**`mamba` (Mac)**:\r\n```\r\nmamba create -y -n sleap_v1.4.1a2 -c conda-forge -c anaconda -c sleap\u002Flabel\u002Fdev sleap=1.4.1a2\r\n```\r\n\r\n**`pip` (any OS except Apple Silicon)**:\r\n```\r\npip install sleap[pypi]==1.4.1a2\r\n```\r\n\r\n# What's Changed\r\n\r\n## Fixes\r\n* Fix zmq inference by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1800\r\n\r\n## Workflow Changes\r\n* Fix windows conda package upload and build ci by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1792\r\n* Bump to v1.4.1a2 by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1835\r\n\r\n## Enhancements and Refactors\r\n* Set selected instance to None after removal by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1808\r\n* Add `InstancesList` class to handle backref to `LabeledFrame` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1807\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.4.1a1...v1.4.1a2","2024-06-28T18:40:18",{"id":209,"version":210,"summary_zh":211,"released_at":212},103822,"v1.4.1a1","SLEAP v1.4.1a1 is a pre-release. See [1.3.3](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.3.3) for the latest stable release. There are many changes to dependencies in this pre-release: if you are having installation issues with v1.3.3, you should try this version instead.\r\n\r\n# Quick install\r\n**`mamba` (Windows\u002FLinux\u002FGPU)**:\r\n```\r\nmamba create -y -n sleap_v1.4.1a1 -c conda-forge -c nvidia -c sleap\u002Flabel\u002Fdev -c anaconda sleap=1.4.1a1\r\n```\r\n\r\n**`mamba` (Mac)**:\r\n```\r\nmamba create -y -n sleap_v1.4.1a1 -c conda-forge -c anaconda -c sleap\u002Flabel\u002Fdev sleap=1.4.1a1\r\n```\r\n\r\n**`pip` (any OS except Apple Silicon)**:\r\n```\r\npip install sleap[pypi]==1.4.1a1\r\n```\r\n\r\n# What's Changed\r\n\r\n## Enhancements\r\n* Add options to set background color when exporting video by @scott-yj-yang in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1328\r\n* Increase range on batch size by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1513\r\n* Add resize\u002Fscroll to training GUI by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1565\r\n* support loading slp files with non-compound types and str in metadata by @lambdaloop in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1566\r\n* change inference pipeline option to tracking-only by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1666\r\n* Only propagate Transpose Tracks when propagate is checked by @vaibhavtrip29 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1748\r\n* Add batch size to GUI for inference by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1771\r\n* Add ZMQ support via GUI and CLI by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1780\r\n* Change menu name to match deleting predictions beyond max instance by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1790\r\n* Adding ragged metadata to `info.json` by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1765\r\n* Add option to export to CSV via sleap-convert and API by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1730\r\n\r\n## Refactors\r\n* Set default callable for `match_lists_function` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1520\r\n* Allow passing in `Labels` to `app.main` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1524\r\n* Replace (broken) `--unrag` with `--ragged` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1539\r\n* Add function to create app by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1546\r\n* Refactor `AddInstance` command by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1561\r\n\r\n\r\n## Fixes\r\n* Graceful failing with seeking errors by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1712\r\n* Fix IndexError for hdf5 file import for single instance analysis files by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1695\r\n* Import DLC with uniquebodyparts, add Tracks by @getzze in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1562\r\n* Make the hdf5 videos store as int8 format by @lambdaloop in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1559\r\n* Scale new instances to new frame size by @ssrinath22 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1568\r\n* Fix package export by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1619\r\n* View Hyperparameter nonetype fix by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1766\r\n\r\n## Dependency Changes\r\n* Replace imgaug with albumentations by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1623\r\n* Fix out of bounds albumentations issues and update dependencies by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1724\r\n* Update to new TensorFlow conda package by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1726\r\n* Fix conda builds by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1776\r\n\r\n## Workflow Changes\r\n* Fix CI on macosx-arm64 by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1734\r\n* Upgrade build actions for release by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1779\r\n* Fix website build and remove build cache across workflows by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1786\r\n* Bump to 1.4.1a1 by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1791\r\n\r\n## Website Changes\r\n* Add ABL:AOC 2023 Workshop link by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1673\r\n\r\n## New Contributors\r\n* @scott-yj-yang made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1328\r\n* @lambdaloop made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1559\r\n* @ssrinath22 made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1568\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.3.3...v1.4.1a1","2024-06-03T02:53:05",{"id":214,"version":215,"summary_zh":216,"released_at":217},103823,"v1.4.1a0","## What's Changed\r\n* Add options to set background color when exporting video by @scott-yj-yang in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1328\r\n* Increase range on batch size by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1513\r\n* Set default callable for `match_lists_function` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1520\r\n* Allow passing in `Labels` to `app.main` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1524\r\n* Replace (broken) `--unrag` with `--ragged` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1539\r\n* Add function to create app by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1546\r\n* Refactor `AddInstance` command by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1561\r\n* Import DLC with uniquebodyparts, add Tracks by @getzze in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1562\r\n* Make the hdf5 videos store as int8 format by @lambdaloop in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1559\r\n* Scale new instances to new frame size by @ssrinath22 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1568\r\n* Fix package export by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1619\r\n* Add resize\u002Fscroll to training GUI by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1565\r\n* support loading slp files with non-compound types and str in metadata by @lambdaloop in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1566\r\n* change inference pipeline option to tracking-only by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1666\r\n* Add ABL:AOC 2023 Workshop link by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1673\r\n* Graceful failing with seeking errors by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1712\r\n* Fix IndexError for hdf5 file import for single instance analysis files by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1695\r\n* Replace imgaug with albumentations by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1623\r\n* Fix out of bounds albumentations issues and update dependencies by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1724\r\n* Update to new TensorFlow conda package by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1726\r\n* Fix CI on macosx-arm64 by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1734\r\n* Add option to export to CSV via sleap-convert and API by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1730\r\n* Only propagate Transpose Tracks when propagate is checked by @vaibhavtrip29 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1748\r\n* View Hyperparameter nonetype fix by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1766\r\n* Adding ragged metadata to `info.json` by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1765\r\n* Add batch size to GUI for inference by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1771\r\n* Fix conda builds by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1776\r\n* Upgrade build actions for release by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1779\r\n\r\n## New Contributors\r\n* @scott-yj-yang made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1328\r\n* @lambdaloop made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1559\r\n* @ssrinath22 made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1568\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.3.3...v1.4.1a0","2024-05-29T01:00:09",{"id":219,"version":220,"summary_zh":221,"released_at":222},103824,"v1.4.0a0","## What's Changed\r\n* Add options to set background color when exporting video by @scott-yj-yang in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1328\r\n* Increase range on batch size by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1513\r\n* Set default callable for `match_lists_function` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1520\r\n* Allow passing in `Labels` to `app.main` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1524\r\n* Replace (broken) `--unrag` with `--ragged` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1539\r\n* Add function to create app by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1546\r\n* Refactor `AddInstance` command by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1561\r\n* Import DLC with uniquebodyparts, add Tracks by @getzze in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1562\r\n* Make the hdf5 videos store as int8 format by @lambdaloop in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1559\r\n* Scale new instances to new frame size by @ssrinath22 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1568\r\n* Fix package export by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1619\r\n* Add resize\u002Fscroll to training GUI by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1565\r\n* support loading slp files with non-compound types and str in metadata by @lambdaloop in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1566\r\n* change inference pipeline option to tracking-only by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1666\r\n* Add ABL:AOC 2023 Workshop link by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1673\r\n* Graceful failing with seeking errors by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1712\r\n* Fix IndexError for hdf5 file import for single instance analysis files by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1695\r\n* Replace imgaug with albumentations by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1623\r\n* Fix out of bounds albumentations issues and update dependencies by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1724\r\n* Update to new TensorFlow conda package by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1726\r\n* Fix CI on macosx-arm64 by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1734\r\n* Add option to export to CSV via sleap-convert and API by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1730\r\n* Only propagate Transpose Tracks when propagate is checked by @vaibhavtrip29 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1748\r\n* View Hyperparameter nonetype fix by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1766\r\n* Adding ragged metadata to `info.json` by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1765\r\n* Add batch size to GUI for inference by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1771\r\n* Fix conda builds by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1776\r\n\r\n## New Contributors\r\n* @scott-yj-yang made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1328\r\n* @lambdaloop made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1559\r\n* @ssrinath22 made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1568\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.3.3...v1.4.0a0","2024-05-20T18:11:14",{"id":224,"version":225,"summary_zh":226,"released_at":227},103825,"v1.3.3","This is a brown-bag release following insufficient restrictions on allowable `tensorflow` versions for the \"pypi\" extra `sleap[pypi]` in 1.3.2. While the conda packages for 1.3.2 were not affected (since `tensorflow` is pulled in from anaconda), the PyPI only package installed via `pip install sleap[pypi]` had conflicts between the version of `tensorflow` and the version of `keras`. See [1.3.0](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.3.0), [1.3.1](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.3.1), and  [1.3.2](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.3.2) for previous notable changes. \r\n\r\n#### From 1.3.2+, to install SLEAP through pip use `pip install sleap[pypi]` to ensure all dependencies are gathered.\r\n\r\nAs a reminder:\r\n\r\n> The 1.3.1 dependency update requires [Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Findex.html) for faster dependency resolution. If you already have anaconda installed, then you _can_ set the solver to libmamba in the base environment:\r\n>```\r\n>conda update -n base conda\r\n>conda install -n base conda-libmamba-solver\r\n>conda config --set solver libmamba\r\n>```\r\n>Any subsequent `mamba` commands in the docs will need to be replaced with `conda` if you choose to use your existing Anaconda installation. \r\n>\r\n>Otherwise, follow the [recommended installation instruction for Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Finstallation.html).\r\n\r\n# Quick install\r\n**`mamba` (Windows\u002FLinux\u002FGPU)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c nvidia -c sleap -c anaconda sleap=1.3.3\r\n```\r\n\r\n**`mamba` (Mac)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c anaconda -c sleap sleap=1.3.3\r\n```\r\n\r\n**`pip` (any OS except Apple Silicon)**:\r\n```\r\npip install sleap[pypi]==1.3.3\r\n```\r\n\r\n# Full Changelog\r\n\r\n## Fixes\r\n* Do not try to remove item if already deleted by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1498\r\n* Set `LD_LIBRARY_PATH` on mamba activate by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1496\r\n* Reset `LD_LIBRARY_PATH` on deactivate by @roomrys in #1502\r\n\r\n## Dependencies\r\n* Add version restirctions to tensorflow for pypi by @roomrys in #1485\r\n* Remove `imageio` pin by @roomrys in #1501\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.3.2...v1.3.3","2023-09-15T23:54:41",{"id":229,"version":230,"summary_zh":231,"released_at":232},103826,"v1.3.2","SLEAP 1.3.2 adds some nice usability features thanks to both the community ideas and new contributors! See [1.3.0](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.3.0) and [1.3.1](https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Freleases\u002Ftag\u002Fv1.3.1) for previous notable changes. \r\n\r\n#### From 1.3.2+, to install SLEAP through PyPI use `pip install sleap[pypi]` to ensure all dependencies are gathered.\r\n\r\nAs a reminder:\r\n\r\n> The 1.3.1 dependency update requires [Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Findex.html) for faster dependency resolution. If you already have anaconda installed, then you _can_ set the solver to libmamba in the base environment:\r\n>```\r\n>conda update -n base conda\r\n>conda install -n base conda-libmamba-solver\r\n>conda config --set solver libmamba\r\n>```\r\n>Any subsequent `mamba` commands in the docs will need to be replaced with `conda` if you choose to use your existing Anaconda installation. \r\n>\r\n>Otherwise, follow the [recommended installation instruction for Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Finstallation.html).\r\n\r\n# Quick install\r\n**`mamba` (Windows\u002FLinux\u002FGPU)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c nvidia -c sleap -c anaconda sleap=1.3.2\r\n```\r\n\r\n**`mamba` (Mac)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c anaconda -c sleap sleap=1.3.2\r\n```\r\n\r\n**`pip` (any OS except Apple Silicon)**:\r\n```\r\npip install sleap[pypi]==1.3.2\r\n```\r\n\r\n# Highlights\r\n* Limit max tracks via track-local queues by @shrivaths16 and @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1447\r\n* Add option to remove videos in batch by @gitttt-1234 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1382 and https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1406\r\n* Add shortcut to export analysis for current video by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1414 and https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1444\r\n* Add video path and frame indices to metrics by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1396\r\n* Add a button for copying model config to clipboard by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1433\r\n* Add Option to Export CSV by @gitttt-1234 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1438\r\n\r\n# Full Changelog\r\n\r\n## Enhancements\r\n* Add option to remove videos in batch by @gitttt-1234 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1382 and https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1406\r\n* Add `Track` when add `Instance` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1408\r\n* Add `Video` to cache when adding `Track` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1407\r\n* Add shortcut to export analysis for current video by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1414 and https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1444\r\n* Add video path and frame indices to metrics by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1396\r\n* Improve error message for detecting video backend by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1441\r\n* Add a button for copying model config to clipboard by @KevinZ0217 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1433\r\n* Add Option to Export CSV by @gitttt-1234 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1438\r\n* Limit max tracks via track-local queues by @shrivaths16 and @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1447\r\n\r\n## Fixes\r\n* Minor fix in computation of OKS by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1383 and https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1399\r\n* Fix `Filedialog` to work across (mac)OS by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1393\r\n* Fix panning bounding box by @gitttt-1234 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1398\r\n* Fix skeleton templates by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1404\r\n* Fix labels export for json by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1410\r\n* Correct GUI state emulation by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1422\r\n* Update status message on status bar by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1411\r\n* Fix error thrown when last video is deleted  by @shrivaths16 in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1421\r\n* Add model folder to the unzip path by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1445\r\n* Fix drag and drop by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1449\r\n\r\n## Dependencies\r\n* Pin micromamba version by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1376\r\n* Add pip extras by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1481\r\n\r\n## New Contributors\r\n* @shrivaths16 made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1383\r\n* @gitttt-1234 made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1382\r\n* @KevinZ0217 made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1414\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.3.1...v1.3.2","2023-09-10T17:07:01",{"id":234,"version":235,"summary_zh":236,"released_at":237},103827,"v1.3.1","After the massive 1.3.0 release, SLEAP 1.3.1 underwent a much needed dependency and build update. SLEAP 1.3.1 has conda packages for Mac OS X and Apple Silicon :tada:. In terms of features, 1.3.1 has just a few small upgrades\u002Ffixes. Be sure to check back in for bigger features still in the works! :construction: :hammer: 👀\r\n\r\n> The 1.3.1 dependency update requires [Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Findex.html) for faster dependency resolution. If you already have anaconda installed, then you can install Mamba in the base environment:\r\n>```\r\n>conda install mamba -n base -c conda-forge\r\n>```\r\n>Otherwise, follow the [recommended installation instruction for Mamba](https:\u002F\u002Fmamba.readthedocs.io\u002Fen\u002Flatest\u002Finstallation.html).\r\n\r\n# Quick install\r\n**`mamba` (Windows\u002FLinux\u002FGPU)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c nvidia -c sleap -c anaconda sleap=1.3.1\r\n```\r\n\r\n**`mamba` (Mac)**:\r\n```\r\nmamba create -y -n sleap -c conda-forge -c anaconda -c sleap sleap=1.3.1\r\n```\r\n\r\n**`pip` (any OS except Apple Silicon)**:\r\n```\r\npip install sleap==1.3.1\r\n```\r\n\r\n# Highlights\r\n* Update environment creation by @roomrys in #1366\r\n* Add `--max_instances` to `sleap-track` and GUI by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1305\r\n* Increase GUI crop size range from 512 to 832 by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1295\r\n* Allow returning PAF graph during low level inference by @calebweinreb in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1329\r\n* Fix GUI resume training by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1314\r\n* Fixes GPU memory polling using environment variable filtering by @ericleonardis in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1272\r\n\r\n# Full Changelog\r\n\r\n## Enhancements\r\n* Centralize video extensions by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1244\r\n* Organize docks by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1265\r\n* Increase GUI crop size range from 512 to 832 by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1295\r\n* Add `--max_instances` to `sleap-track` and GUI by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1305\r\n* Allow returning PAF graph during low level inference by @calebweinreb in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1329\r\n\r\n## Fixes\r\n* Disable data caching by default for SingleImageVideos by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1243\r\n* Fix single frame GUI increment by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1254\r\n* Fix conversion to numpy array when last frame(s) do not have labels by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1307\r\n* Ensure frames to predict list is unique by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1293\r\n* Fix GUI resume training by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1314\r\n* Do not choose `top_k` instances if `max_instances` \u003C num centroids by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1313\r\n* Remove `--labels` and redundant `data_path` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1326\r\n* Create copy of config info to modify (gui) by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1325\r\n* Fixes GPU memory polling using environment variable filtering by @ericleonardis in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1272\r\n* Set `split_by_inds`, `test_labels`, and `validation_labels` to default (GUI) by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1331\r\n* Fix (remove) `SingleImageVideo` caching by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1330\r\n\r\n## Dependencies\r\n* Update environment creation by @roomrys in #1366\r\n\r\n## New Contributors\r\n* @ericleonardis made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1272\r\n* @calebweinreb made their first contribution in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1329\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fcompare\u002Fv1.3.0...v1.3.1","2023-06-30T20:43:59",{"id":239,"version":240,"summary_zh":241,"released_at":242},103828,"v1.3.0","For 1.3.0 we want to give our users some cool new features worth upgrading for! This release includes a many enhancements we hope our users will enjoy as well as its fair share of bug fixes.\r\n\r\n# Quick install\r\n**`conda` (Windows\u002FLinux\u002FGPU)**:\r\n```\r\nconda create -y -n sleap -c sleap -c nvidia -c conda-forge sleap=1.3.0\r\n```\r\n\r\n**`pip` (any OS except Apple Silicon)**:\r\n```\r\npip install sleap==1.3.0\r\n```\r\n\r\n## What's Changed\r\n\r\n# Highlights\r\n* Added scaling functionality for both the instances and bounding box.  by @sean-afshar in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1133\r\n* Add Skeleton Templates by @aaprasad in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1122\r\n* Resumable Training by @jimzers in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1130\r\n* Tracking: robust assignment of the best score to an instance by @getzze in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1062\r\n* Set max instances for top down models by @sheridana in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1070\r\n* Flexibly resize input layer of `tf.keras.Model` upon loading trained model by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1084\r\n* GUI Training: Use hidden params from loaded config by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1053\r\n* Nix export of tracking results by @jgrewe in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1068\r\n* Expose MoveNet to the Inference GUI by @sheridana in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1190\r\n* Add option to \"Add Videos...\" from single image files by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1183\r\n* Add GUI and API option to remove unused tracks by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1210\r\n* Add instance and track copy\u002Fpasting by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1206\r\n* Expose supervised ID models to GUI by @sheridana in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1213\r\n* Ensure data format compatibility by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1222\r\n\r\n# Full Changelog\r\n\r\n## Documentation\r\n* Change 'M1' to 'Apple Silicon' by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1188\r\n* Update the Documentation badge by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1211\r\n\r\n## Enhancements\r\n* GUI Training: Use hidden params from loaded config by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1053\r\n* Add optional unragging arg to model export by @sheridana in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1054\r\n* Tracking: robust assignment of the best score to an instance by @getzze in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1062\r\n* Set max instances for top down models by @sheridana in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1070\r\n* Flexibly resize input layer of `tf.keras.Model` upon loading trained model by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1084\r\n* Add Option to Make Trail Shade Darker\u002FLighter  by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1103\r\n* Nix export of tracking results by @jgrewe in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1068\r\n* Added scaling functionality for both the instances and bounding box.  by @sean-afshar in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1133\r\n* Add better error message for top down by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1121\r\n* Add central padding to SizeMatcher by @jiayinghsu in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1129\r\n* Added MoveNet as an external model reference by @jiayinghsu in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1141\r\n* Resumable Training by @jimzers in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1130\r\n* GenericTableModel\u002FView improvements by @jgrewe in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1163\r\n* Add Skeleton Templates by @aaprasad in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1122\r\n* Add better error message for top down by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1121\r\n* Add option of 2 for marker size by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1205\r\n* Support new DLC multi-animal configs by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1204\r\n* Expose MoveNet to the Inference GUI by @sheridana in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1190\r\n* Add option to \"Add Videos...\" from single image files by @eberrigan in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1183\r\n* Add GUI and API option to remove unused tracks by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1210\r\n* Add instance and track copy\u002Fpasting by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1206\r\n* Expose supervised ID models to GUI by @sheridana in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1213\r\n* Toggle grayscale of all videos using \"Toggle Grayscale\" button by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1215\r\n\r\n## Fixes\r\n* Fix config option to `split_by_inds` by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1060\r\n* Don't create instances during inference if no points were found by @talmo in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1073\r\n* Add one-line fix to VideoWriterSkyvideo by @roomrys in https:\u002F\u002Fgithub.com\u002Ftalmolab\u002Fsleap\u002Fpull\u002F1082\r\n* Fix parser for sleap-export by @roomrys in https:\u002F\u002Fgit","2023-03-23T22:11:50"]