[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-ourownstory--neural_prophet":3,"tool-ourownstory--neural_prophet":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},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,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},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",160015,2,"2026-04-18T11:30:52",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},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。",109154,"2026-04-18T11:18:24",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},6121,"gemini-cli","google-gemini\u002Fgemini-cli","gemini-cli 是一款由谷歌推出的开源 AI 命令行工具，它将强大的 Gemini 大模型能力直接集成到用户的终端环境中。对于习惯在命令行工作的开发者而言，它提供了一条从输入提示词到获取模型响应的最短路径，无需切换窗口即可享受智能辅助。\n\n这款工具主要解决了开发过程中频繁上下文切换的痛点，让用户能在熟悉的终端界面内直接完成代码理解、生成、调试以及自动化运维任务。无论是查询大型代码库、根据草图生成应用，还是执行复杂的 Git 操作，gemini-cli 都能通过自然语言指令高效处理。\n\n它特别适合广大软件工程师、DevOps 人员及技术研究人员使用。其核心亮点包括支持高达 100 万 token 的超长上下文窗口，具备出色的逻辑推理能力；内置 Google 搜索、文件操作及 Shell 命令执行等实用工具；更独特的是，它支持 MCP（模型上下文协议），允许用户灵活扩展自定义集成，连接如图像生成等外部能力。此外，个人谷歌账号即可享受免费的额度支持，且项目基于 Apache 2.0 协议完全开源，是提升终端工作效率的理想助手。",100752,"2026-04-10T01:20:03",[52,13,15,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":32,"last_commit_at":59,"category_tags":60,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":73,"owner_avatar_url":74,"owner_bio":75,"owner_company":76,"owner_location":77,"owner_email":75,"owner_twitter":75,"owner_website":75,"owner_url":78,"languages":79,"stars":88,"forks":89,"last_commit_at":90,"license":91,"difficulty_score":32,"env_os":92,"env_gpu":93,"env_ram":93,"env_deps":94,"category_tags":103,"github_topics":104,"view_count":32,"oss_zip_url":75,"oss_zip_packed_at":75,"status":17,"created_at":125,"updated_at":126,"faqs":127,"releases":157},9159,"ourownstory\u002Fneural_prophet","neural_prophet","NeuralProphet: A simple forecasting package","NeuralProphet 是一个基于 PyTorch 构建的开源时间序列预测框架，旨在让复杂的神经网络预测变得简单易懂且具备可解释性。它巧妙融合了神经网络的强大学习能力与传统时间序列算法的稳定性，灵感源自 Facebook Prophet 和 AR-Net，专为解决高频（如亚日级）且长周期数据的预测难题而生。\n\n面对传统黑盒模型难以调试、普通用户上手门槛高的问题，NeuralProphet 提供了“人在回路”的迭代式建模体验。用户只需几行代码即可定义、定制、可视化并评估模型，能够快速建立基准模型，通过直观解读结果来持续优化，直到获得满意的预测效果。虽然其开箱即用的精度可能并非极致，但其高度的可定制性和透明度让用户能深入理解数据背后的规律。\n\n这款工具非常适合需要处理复杂时间序列数据的开发者、数据科学家及研究人员使用，尤其是那些希望在不牺牲模型可解释性的前提下利用深度学习技术的团队。无论是进行销售趋势分析、能源消耗预测还是其他周期性业务规划，NeuralProphet 都能提供一个灵活、友好且专业的解决方案，帮助用户轻松掌控预测流程。","[![GitHub release (latest SemVer)](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fv\u002Frelease\u002Fourownstory\u002Fneural_prophet?logo=github)](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Freleases)\n[![Pypi_Version](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fneuralprophet.svg)](https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fneuralprophet)\n[![Python Version](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.9+-blue?logo=python)](https:\u002F\u002Fwww.python.org\u002F)\n[![Code style: black](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fcode%20style-black-000000.svg)](https:\u002F\u002Fgithub.com\u002Fpsf\u002Fblack)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-brightgreen)](https:\u002F\u002Fopensource.org\u002Flicenses\u002FMIT)\n[![Tests](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Factions\u002Fworkflows\u002Ftests.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Factions\u002Fworkflows\u002Ftests.yml)\n[![codecov](https:\u002F\u002Fcodecov.io\u002Fgh\u002Fourownstory\u002Fneural_prophet\u002Fbranch\u002Fmaster\u002Fgraph\u002Fbadge.svg?token=U5KXCL55DW)](https:\u002F\u002Fcodecov.io\u002Fgh\u002Fourownstory\u002Fneural_prophet)\n[![Slack](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fslack-@neuralprophet-CF0E5B.svg?logo=slack&logoColor=white&labelColor=3F0E40)](https:\u002F\u002Fneuralprophet.slack.com\u002Fjoin\u002Fshared_invite\u002Fzt-sgme2rw3-3dCH3YJ_wgg01IXHoYaeCg#\u002Fshared-invite\u002Femail)\n[![Downloads](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fourownstory_neural_prophet_readme_8e91e2e5c473.png)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fneuralprophet)\n\n![NP-logo-wide_cut](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fourownstory_neural_prophet_readme_5bd25714c10a.png)\n\n\nPlease note that the project is still in beta phase. Please report any issues you encounter or suggestions you have. We will do our best to address them quickly. Contributions are very welcome!\n\n# NeuralProphet: human-centered forecasting\nNeuralProphet is an easy to learn framework for interpretable time series forecasting.\nNeuralProphet is built on PyTorch and combines Neural Networks and traditional time-series algorithms, inspired by [Facebook Prophet](https:\u002F\u002Fgithub.com\u002Ffacebook\u002Fprophet) and [AR-Net](https:\u002F\u002Fgithub.com\u002Fourownstory\u002FAR-Net).\n- With a few lines of code, you can define, customize, visualize, and evaluate your own forecasting models.\n- It is designed for iterative human-in-the-loop model building. That means that you can build a first model quickly, interpret the results, improve, repeat. Due to the focus on interpretability and customization-ability, NeuralProphet may not be the most accurate model out-of-the-box; so, don't hesitate to adjust and iterate until you like your results.\n- NeuralProphet is best suited for time series data that is of higher-frequency (sub-daily) and longer duration (at least two full periods\u002Fyears).\n\n\n## Documentation\nThe [documentation page](https:\u002F\u002Fneuralprophet.com) may not be entirely up to date. Docstrings should be reliable, please refer to those when in doubt. We are working on an improved documentation. We appreciate any help to improve and update the docs.\n\nFor a visual introduction to NeuralProphet, [view this presentation](notes\u002FNeuralProphet_Introduction.pdf).\n\n## Contribute\nWe compiled a [Contributing to NeuralProphet](CONTRIBUTING.md) page with practical instructions and further resources to help you become part of the family. \n\n## Community\n#### Discussion and Help\nIf you have any questions or suggestion, you can participate in [our community right here on Github](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fdiscussions)\n\n#### Slack Chat\nWe also have an active [Slack community](https:\u002F\u002Fjoin.slack.com\u002Ft\u002Fneuralprophet\u002Fshared_invite\u002Fzt-sgme2rw3-3dCH3YJ_wgg01IXHoYaeCg). Come and join the conversation!\n\n## Tutorials\n[![Open All Collab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fourownstory\u002Fneural_prophet)\n\nThere are several [example notebooks](docs\u002Fsource\u002Ftutorials) to help you get started. \n\nYou can find the datasets used in the tutorials, including data preprocessing examples, in our [neuralprophet-data repository](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneuralprophet-data).\n\nPlease refer to our [documentation page](https:\u002F\u002Fneuralprophet.com) for more resources.\n\n### Minimal example\n```python\nfrom neuralprophet import NeuralProphet\n```\nAfter importing the package, you can use NeuralProphet in your code:\n```python\nm = NeuralProphet()\nmetrics = m.fit(df)\nforecast = m.predict(df)\n```\nYou can visualize your results with the inbuilt plotting functions:\n```python\nfig_forecast = m.plot(forecast)\nfig_components = m.plot_components(forecast)\nfig_model = m.plot_parameters()\n```\nIf you want to forecast into the unknown future, extend the dataframe before predicting:\n```python\nm = NeuralProphet().fit(df, freq=\"D\")\ndf_future = m.make_future_dataframe(df, periods=30)\nforecast = m.predict(df_future)\nfig_forecast = m.plot(forecast)\n```\n## Install\nYou can now install neuralprophet directly with pip:\n```shell\npip install neuralprophet\n```\n\n### Install options\n\nIf you plan to use the package in a Jupyter notebook, we recommended to install the 'live' version:\n```shell\npip install neuralprophet[live]\n```\nThis will allow you to enable `plot_live_loss` in the `fit` function to get a live plot of train (and validation) loss.\n\nIf you would like the most up to date version, you can instead install directly from github:\n```shell\ngit clone \u003Ccopied link from github>\ncd neural_prophet\npip install .\n```\n\nNote for Windows users: Please use WSL2.\n\n## Features\n### Model components\n* Autoregression: Autocorrelation modelling - linear or NN (AR-Net).\n* Trend: Piecewise linear trend with optional automatic changepoint detection.\n* Seasonality: Fourier terms at different periods such as yearly, daily, weekly, hourly.\n* Lagged regressors: Lagged observations (e.g temperature sensor) - linear or NN.\n* Future regressors: In advance known features (e.g. temperature forecast) - linear or NN.\n* Events: Country holidays & recurring custom events.\n* Global Modeling: Components can be local, global or 'glocal' (global + regularized local)\n\n\n### Framework features\n* Multiple time series: Fit a global\u002Fglocal model with (partially) shared model parameters.\n* Uncertainty: Estimate values of specific quantiles - Quantile Regression.\n* Regularize modelling components.\n* Plotting of forecast components, model coefficients and more.\n* Time series crossvalidation utility.\n* Model checkpointing and validation.\n\n\n### Coming soon\u003Csup>:tm:\u003C\u002Fsup>\n\n* Cross-relation of lagged regressors.\n* Static metadata regression for multiple series\n* Logistic growth for trend component.\n\nFor a list of past changes, please refer to the [releases page](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Freleases).\n\n## Cite\nPlease cite [NeuralProphet](https:\u002F\u002Farxiv.org\u002Fabs\u002F2111.15397) in your publications if it helps your research:\n```\n@misc{triebe2021neuralprophet,\n      title={NeuralProphet: Explainable Forecasting at Scale}, \n      author={Oskar Triebe and Hansika Hewamalage and Polina Pilyugina and Nikolay Laptev and Christoph Bergmeir and Ram Rajagopal},\n      year={2021},\n      eprint={2111.15397},\n      archivePrefix={arXiv},\n      primaryClass={cs.LG}\n}\n```\n### Many Thanks To Our Contributors:\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fgraphs\u002Fcontributors\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fourownstory_neural_prophet_readme_8200be501cf3.png\" \u002F>\n\u003C\u002Fa>\n\n## About\nNeuralProphet is an open-source community project, supported by awesome people like you. \nIf you are interested in joining the project, please feel free to reach out to me (Oskar) - you can find my email on the [NeuralProphet Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2111.15397).\n","[![GitHub release (latest SemVer)](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fv\u002Frelease\u002Fourownstory\u002Fneural_prophet?logo=github)](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Freleases)\n[![Pypi_Version](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fneuralprophet.svg)](https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fneuralprophet)\n[![Python Version](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.9+-blue?logo=python)](https:\u002F\u002Fwww.python.org\u002F)\n[![Code style: black](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fcode%20style-black-000000.svg)](https:\u002F\u002Fgithub.com\u002Fpsf\u002Fblack)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-brightgreen)](https:\u002F\u002Fopensource.org\u002Flicenses\u002FMIT)\n[![Tests](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Factions\u002Fworkflows\u002Ftests.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Factions\u002Fworkflows\u002Ftests.yml)\n[![codecov](https:\u002F\u002Fcodecov.io\u002Fgh\u002Fourownstory\u002Fneural_prophet\u002Fbranch\u002Fmaster\u002Fgraph\u002Fbadge.svg?token=U5KXCL55DW)](https:\u002F\u002Fcodecov.io\u002Fgh\u002Fourownstory\u002Fneural_prophet)\n[![Slack](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fslack-@neuralprophet-CF0E5B.svg?logo=slack&logoColor=white&labelColor=3F0E40)](https:\u002F\u002Fneuralprophet.slack.com\u002Fjoin\u002Fshared_invite\u002Fzt-sgme2rw3-3dCH3YJ_wgg01IXHoYaeCg#\u002Fshared-invite\u002Femail)\n[![Downloads](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fourownstory_neural_prophet_readme_8e91e2e5c473.png)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fneuralprophet)\n\n![NP-logo-wide_cut](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fourownstory_neural_prophet_readme_5bd25714c10a.png)\n\n\n请注意，该项目目前仍处于测试阶段。如果您遇到任何问题或有任何建议，请随时报告。我们将尽最大努力尽快解决。非常欢迎您的贡献！\n\n# NeuralProphet：以人为本的预测\nNeuralProphet 是一个易于学习、可解释的时间序列预测框架。\nNeuralProphet 基于 PyTorch 构建，结合了神经网络和传统的时间序列算法，灵感来源于 [Facebook Prophet](https:\u002F\u002Fgithub.com\u002Ffacebook\u002Fprophet) 和 [AR-Net](https:\u002F\u002Fgithub.com\u002Fourownstory\u002FAR-Net)。\n- 只需几行代码，您就可以定义、自定义、可视化并评估您自己的预测模型。\n- 它专为迭代的人机交互式建模而设计。这意味着您可以快速构建第一个模型，解读结果，改进并重复这一过程。由于注重可解释性和可定制性，NeuralProphet 可能不是开箱即用最精确的模型；因此，请毫不犹豫地调整和迭代，直到您对结果满意为止。\n- NeuralProphet 最适合高频（日以下）且较长时长（至少两个完整周期\u002F年）的时间序列数据。\n\n\n## 文档\n[文档页面](https:\u002F\u002Fneuralprophet.com)可能尚未完全更新。在有疑问时，请参考可靠的 docstring。我们正在努力改进文档，并欢迎任何帮助来完善和更新文档。\n\n如需了解 NeuralProphet 的直观介绍，请查看此演示文稿：[NeuralProphet_介绍.pdf](notes\u002FNeuralProphet_Introduction.pdf)。\n\n## 贡献\n我们整理了一份 [贡献指南](CONTRIBUTING.md)，其中包含实用说明和其他资源，以帮助您加入我们的社区。\n\n## 社区\n#### 讨论与帮助\n如果您有任何问题或建议，可以参与我们在 [Github 上的社区讨论](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fdiscussions)。\n\n#### Slack 聊天\n我们还有一个活跃的 [Slack 社区](https:\u002F\u002Fjoin.slack.com\u002Ft\u002Fneuralprophet\u002Fshared_invite\u002Fzt-sgme2rw3-3dCH3YJ_wgg01IXHoYaeCg)。欢迎加入我们的对话！\n\n## 教程\n[![Open All Collab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fourownstory\u002Fneural_prophet)\n\n我们提供了几个 [示例笔记本](docs\u002Fsource\u002Ftutorials)，帮助您入门。\n\n教程中使用的数据集，包括数据预处理示例，可以在我们的 [neuralprophet-data 仓库](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneuralprophet-data) 中找到。\n\n更多资源请参阅我们的 [文档页面](https:\u002F\u002Fneuralprophet.com)。\n\n### 最小示例\n```python\nfrom neuralprophet import NeuralProphet\n```\n导入包后，您可以在代码中使用 NeuralProphet：\n```python\nm = NeuralProphet()\nmetrics = m.fit(df)\nforecast = m.predict(df)\n```\n您可以使用内置的绘图函数来可视化结果：\n```python\nfig_forecast = m.plot(forecast)\nfig_components = m.plot_components(forecast)\nfig_model = m.plot_parameters()\n```\n如果您想对未来未知的时间进行预测，在预测前扩展数据框即可：\n```python\nm = NeuralProphet().fit(df, freq=\"D\")\ndf_future = m.make_future_dataframe(df, periods=30)\nforecast = m.predict(df_future)\nfig_forecast = m.plot(forecast)\n```\n## 安装\n现在您可以直接通过 pip 安装 neuralprophet：\n```shell\npip install neuralprophet\n```\n\n### 安装选项\n\n如果您计划在 Jupyter Notebook 中使用该包，我们建议安装“live”版本：\n```shell\npip install neuralprophet[live]\n```\n这样您就可以在 `fit` 函数中启用 `plot_live_loss`，实时查看训练（和验证）损失曲线。\n\n如果您希望使用最新版本，也可以直接从 GitHub 安装：\n```shell\ngit clone \u003C从 GitHub 复制的链接>\ncd neural_prophet\npip install .\n```\n\nWindows 用户请注意：请使用 WSL2。\n\n## 特性\n### 模型组件\n* 自回归：自相关建模——线性或神经网络（AR-Net）。\n* 趋势：分段线性趋势，可选自动变点检测。\n* 季节性：不同周期的傅里叶项，如年度、每日、每周、每小时。\n* 滞后回归量：滞后观测值（例如温度传感器）——线性或神经网络。\n* 未来回归量：预先已知的特征（例如温度预报）——线性或神经网络。\n* 事件：国家节假日及定期自定义事件。\n* 全局建模：组件可以是局部的、全局的或“全球+本地”的混合形式（glocal）。\n\n\n### 框架特性\n* 多时间序列：拟合具有部分共享模型参数的全局\u002F混合模型。\n* 不确定性：估计特定分位数的值——分位数回归。\n* 对建模组件进行正则化。\n* 绘制预测成分、模型系数等。\n* 时间序列交叉验证工具。\n* 模型检查点保存与验证。\n\n\n### 即将推出\u003Csup>:tm:\u003C\u002Fsup>\n\n* 滞后回归量之间的交叉关系。\n* 多个序列的静态元数据回归。\n* 趋势组件中的逻辑增长。\n\n有关过往变更的列表，请参阅 [发布页面](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Freleases)。\n\n## 引用\n如果您在研究中使用了 NeuralProphet，欢迎您在您的出版物中引用它：\n```\n@misc{triebe2021neuralprophet,\n      title={NeuralProphet: 可解释的大规模预测}, \n      author={Oskar Triebe 和 Hansika Hewamalage、Polina Pilyugina、Nikolay Laptev、Christoph Bergmeir 以及 Ram Rajagopal},\n      year={2021},\n      eprint={2111.15397},\n      archivePrefix={arXiv},\n      primaryClass={cs.LG}\n}\n```\n\n### 衷心感谢所有贡献者：\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fgraphs\u002Fcontributors\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fourownstory_neural_prophet_readme_8200be501cf3.png\" \u002F>\n\u003C\u002Fa>\n\n## 关于我们\nNeuralProphet 是一个开源社区项目，由像您这样优秀的人们共同支持。  \n如果您有兴趣加入该项目，请随时联系我（Oskar）——我的邮箱地址可在 [NeuralProphet 论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2111.15397) 中找到。","# NeuralProphet 快速上手指南\n\nNeuralProphet 是一个基于 PyTorch 构建的可解释时间序列预测框架。它结合了神经网络与传统时间序列算法（灵感源自 Facebook Prophet 和 AR-Net），专为高频（亚日级）且长周期（至少两个完整周期\u002F年）的数据设计，支持通过少量代码实现模型的定义、定制、可视化和评估。\n\n## 环境准备\n\n在开始之前，请确保您的开发环境满足以下要求：\n\n*   **操作系统**：推荐 Linux 或 macOS。\n    *   *注意*：Windows 用户请务必使用 **WSL2** (Windows Subsystem for Linux) 以获得最佳兼容性。\n*   **Python 版本**：3.9 或更高版本 (`python-3.9+`)。\n*   **前置依赖**：项目核心依赖 PyTorch，安装时会自动处理。\n\n## 安装步骤\n\n您可以直接使用 `pip` 进行安装。为了提高下载速度，国内用户推荐使用清华或阿里镜像源。\n\n### 1. 基础安装\n适用于大多数标准使用场景：\n```shell\npip install neuralprophet -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n\n### 2. 交互式安装（推荐 Jupyter 用户）\n如果您计划在 Jupyter Notebook 中使用，并希望通过 `plot_live_loss` 参数实时查看训练损失曲线，请安装 `live` 版本：\n```shell\npip install \"neuralprophet[live]\" -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n\n### 3. 源码安装（获取最新版）\n如果需要体验最新的开发版功能：\n```shell\ngit clone https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet.git\ncd neural_prophet\npip install . -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n\n## 基本使用\n\nNeuralProphet 的设计初衷是让人类能够参与到模型构建的循环中（Human-in-the-loop）。以下是最小化的使用示例。\n\n### 1. 导入与建模\n假设您已经有一个包含时间序列数据的 DataFrame (`df`)，其中包含 `ds` (时间戳) 和 `y` (目标值) 列。\n\n```python\nfrom neuralprophet import NeuralProphet\n\n# 初始化模型\nm = NeuralProphet()\n\n# 拟合数据\nmetrics = m.fit(df)\n\n# 进行预测\nforecast = m.predict(df)\n```\n\n### 2. 结果可视化\n利用内置绘图函数直观展示预测结果、组件分解及模型参数：\n\n```python\n# 绘制预测结果\nfig_forecast = m.plot(forecast)\n\n# 绘制各组件（趋势、季节性等）\nfig_components = m.plot_components(forecast)\n\n# 绘制模型参数\nfig_model = m.plot_parameters()\n```\n\n### 3. 未来预测\n若需预测未来未知时间段的数据，需先扩展数据框：\n\n```python\n# 初始化并拟合，指定频率 (例如 \"D\" 代表天)\nm = NeuralProphet().fit(df, freq=\"D\")\n\n# 创建未来数据框，预测未来 30 个周期\ndf_future = m.make_future_dataframe(df, periods=30)\n\n# 执行预测\nforecast = m.predict(df_future)\n\n# 可视化\nfig_forecast = m.plot(forecast)\n```\n\n> **提示**：NeuralProphet 目前处于 Beta 阶段，旨在通过快速迭代和可解释性来优化模型。如果初始结果未达预期，请尝试调整模型配置并重新训练。","某连锁零售企业的数据团队需要预测未来三个月各门店的每小时客流量，以优化排班和库存管理。\n\n### 没有 neural_prophet 时\n- 面对高频（小时级）且长达两年的复杂数据，传统统计模型难以捕捉非线性趋势，而自定义深度学习模型代码量大、调试困难。\n- 模型如同“黑盒”，业务部门无法理解为何预测周末销量激增，导致信任度低，决策层不敢依据预测结果调整资源。\n- 每次加入新的影响因素（如促销活动或天气变化），都需要重构整个网络结构，迭代周期长达数周，无法响应快速变化的市场需求。\n- 缺乏内置的可解释性可视化工具，分析师需手动编写大量代码来拆解趋势、季节性和事件影响，效率极低。\n\n### 使用 neural_prophet 后\n- 仅需几行代码即可构建结合神经网络与传统时间序列优势的模型，轻松处理高频长周期数据，快速获得基准预测结果。\n- 利用其内置的可解释性组件，直观展示节假日、促销等具体事件对销量的量化贡献，让业务方清晰理解预测逻辑，建立信任。\n- 支持“人在回路”的迭代模式，业务专家可随时添加或调整协变量（如新增门店开业信息），模型即时更新并反馈效果，将迭代周期缩短至小时级。\n- 直接调用可视化接口生成详细的成分分解图，自动呈现趋势、周期性及外部因子的影响，大幅减少手动分析工作量。\n\nneural_prophet 通过平衡深度学习的精度与传统方法的透明度，让非算法专家也能高效构建可解释、可迭代的高频时间序列预测模型。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fourownstory_neural_prophet_7947c5fa.png","ourownstory","Oskar Triebe","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fourownstory_20d04409.png",null,"Stanford University","Palo Alto","https:\u002F\u002Fgithub.com\u002Fourownstory",[80,84],{"name":81,"color":82,"percentage":83},"Python","#3572A5",99.9,{"name":85,"color":86,"percentage":87},"Shell","#89e051",0.1,4266,512,"2026-04-18T03:08:08","MIT","Linux, macOS","未说明",{"notes":95,"python":96,"dependencies":97},"Windows 用户请使用 WSL2 (Windows Subsystem for Linux) 运行。该项目目前处于 Beta 阶段。若在 Jupyter Notebook 中使用实时损失绘图功能，建议安装 'neuralprophet[live]' 版本。该工具最适合高频（亚日级）且持续时间较长（至少两个完整周期\u002F年）的时间序列数据。","3.9+",[98,99,100,101,102],"torch","pandas","numpy","matplotlib","plotly",[14],[105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124],"forecasting","time-series","machine-learning","fbprophet","prophet","forecast","artificial-intelligence","prediction","trend","seasonality","autoregression","pytorch","timeseries","forecasting-algorithm","forecasting-model","neuralprophet","neural","neural-network","python","deep-learning","2026-03-27T02:49:30.150509","2026-04-19T03:09:41.796835",[128,133,138,143,148,152],{"id":129,"question_zh":130,"answer_zh":131,"source_url":132},41129,"为什么通过 Conda 安装 Neural Prophet 时版本不是最新的（例如只安装了 0.3.2 而不是 0.6.2）？","Conda 渠道上的包版本更新可能滞后于 PyPI。如果遇到版本不一致问题，官方建议在 Windows 上使用 WSL2 环境，并优先通过 pip 进行安装以获取最新版本。由于 Conda 维护的延迟，推荐使用 pip install neuralprophet 来确保安装最新版。","https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fissues\u002F1435",{"id":134,"question_zh":135,"answer_zh":136,"source_url":137},41127,"如何在训练时完全禁用或阻止生成 lightning_logs 文件夹？","Neural Prophet 本身没有直接参数来完全禁用日志文件夹的创建。要阻止文件夹生成，需要调整底层的 PyTorch Lightning 配置设置。虽然可以通过修改源码实现，但官方建议保留日志以便调试。如果在云端运行，最佳实践是配置日志将其流式传输到云存储容器（如 S3），而不是试图完全禁用本地日志写入。","https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fissues\u002F1527",{"id":139,"question_zh":140,"answer_zh":141,"source_url":142},41128,"在多线程 GPU 训练时遇到 'Encountered different devices in metric calculation' 错误如何解决？","这是一个已知问题，发生在多线程 GPU 训练中指标计算时张量设备不一致的情况。该问题已在 PR #1365 中修复。请确保您使用的是包含此修复的最新版本代码。如果自行构建，请拉取最新代码并重新安装：git clone \u003C仓库链接> && cd neural_prophet && pip install .。","https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fissues\u002F1361",{"id":144,"question_zh":145,"answer_zh":146,"source_url":147},41125,"升级到新版本（0.9.0）后，为什么模型训练时间变长了？","训练时间的变化可能与 PyTorch Lightning 2.0 的升级有关。如果您发现训练变慢，可以尝试在 model.fit() 方法中添加 minimal=True 参数，这通常能显著加快本地和服务器的训练过程。例如：model.fit(training_data, freq=engine['freq'], epochs=engine['epochs'], minimal=True)。在某些云环境中，该更新实际上可将训练时间减少高达 50%。","https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fissues\u002F1629",{"id":149,"question_zh":150,"answer_zh":151,"source_url":137},41126,"在 Docker 容器中运行时遇到 PermissionError: [Errno 13] Permission denied 错误怎么办？","此问题通常与 PyTorch Lightning 的日志写入权限有关。在升级到 0.9.0 版本（包含 PyTorch Lightning 2.x）后，该问题大多已得到解决。如果仍然遇到此问题，建议检查容器内的文件权限设置，或者考虑将日志流式传输到云存储（如 AWS S3 或 Azure Blob）以避免本地写入冲突。目前无需手动清除 lightning_logs 文件夹即可正常运行。",{"id":153,"question_zh":154,"answer_zh":155,"source_url":156},41130,"项目是否已经迁移支持 PyTorch 2.0？","是的，Neural Prophet 已经完成向 PyTorch 2.0 的迁移。相关工作已在 PR #1404 中完成并合并。用户可以直接使用支持 PyTorch 2.0 的最新版本。","https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fissues\u002F1229",[158,163,168,173,178,183,188,193,198,203,208,213,218,223,228,233,238,243,248,253],{"id":159,"version":160,"summary_zh":161,"released_at":162},324705,"1.0.0rc10","还包括自 Beta 0.9.0 以来的更改。\n\n## 主要更改\n* [重大] 数据加载器：由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1529 中实现的即时表格化功能\n## 其他更改\n* [运维] 由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1596 中优化 future-reg-nn 测试速度\n* [次要] 使测试具有确定性，由 @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1600 中完成\n* [次要] 改进 Season glocal reg 无效参数处理，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1601 中完成\n\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.9.0...1.0.0rc10","2024-06-27T00:00:25",{"id":164,"version":165,"summary_zh":166,"released_at":167},324706,"0.9.0","# 主要变更\n* [DevOps] 升级至 Lightning 2.0，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1514 中完成\n\n## 新特性\n* [次要] 添加节假日细分，由 @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1584 中完成\n* [次要] 支持 io.bytes 输出，由 @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1583 中完成\n\n## 修复\n* [修复] 修复 #1580 漏洞，由 @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1582 中完成\n* [修复] 修复未来回归器问题，由 @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1585 中完成\n* [修复] 修复神经网络回归器形状问题，由 @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1589 中完成\n* [修复] 如果设置了 ar_reg 而 dn_lags 为空，则抛出错误，由 @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1588 中完成\n* [修复] 修复当频率为 NaT 时的推断问题，由 @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1590 中完成\n* [文档] 修复 get_latest_forecast 示例中的拼写错误，由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1551 中完成\n\n## 其他变更\n* [DevOps] 清理 CI 工作流文件及依赖项，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1547 中完成\n* [DevOps] 更新 Poetry，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1587 中完成\n* [DevOps] 运行 black 代码格式化工具，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1593 中完成\n* [DevOps] 将版本号提升至 0.9.0，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1594 中完成\n\n## 新贡献者\n* @MaiBe-ctrl 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1582 中完成了首次贡献\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.8.0...0.9.0","2024-06-21T07:42:22",{"id":169,"version":170,"summary_zh":171,"released_at":172},324707,"0.8.0","注意：与 [1.0.0rc9](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Freleases\u002Ftag\u002F1.0.0rc9) 完全相同\n\n# 主要变更\n* [重大] 全局-局部混合建模 V2：由 @alfonsogarciadecorral 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1008 中实现的局部-全局混合模型\n* [重大] 增加对 Python 3.12 的支持，移除对 Python 3.8 的支持：由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1544 中完成\n\n## 其他变更\n* [修复] 访问 None 类型的 config_refgressors：由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1536 中完成\n* [文档] 将官网更新至 1.0.0rc8：由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1538 中完成\n* [DevOps] 将依赖升级到更新的版本范围：由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1543 中完成\n\n\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.7.0...0.8.0","2024-02-23T20:22:32",{"id":174,"version":175,"summary_zh":176,"released_at":177},324708,"1.0.0rc9","## 变更内容\n* [重大] 添加对 Python 3.12 的支持，移除对 Python 3.8 的支持，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1544 中完成\n* [文档] 将网站更新至 1.0.0rc8 版本，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1538 中完成\n* [开发运维] 将依赖升级到更新的版本范围，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1543 中完成\n\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F1.0.0rc8...1.0.0rc9","2024-02-23T20:18:49",{"id":179,"version":180,"summary_zh":181,"released_at":182},324709,"1.0.0rc8","## 变更内容\n* [修复] 修复了访问 `None` 类型的 `config_refgressors`（由 #1008 引入）的问题，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1536 中完成。\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F1.0.0rc7...1.0.0rc8","2024-02-19T22:05:42",{"id":184,"version":185,"summary_zh":186,"released_at":187},324710,"1.0.0rc7","## 变更内容\n* [重大] 由 @alfonsogarciadecorral 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1008 中提出的 Glocal 建模 v2\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.7.0...1.0.0rc7","2024-02-14T23:49:29",{"id":189,"version":190,"summary_zh":191,"released_at":192},324711,"0.7.0","注意：与 [1.0.0rc6](https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Freleases\u002Ftag\u002F1.0.0rc6) 完全相同  \n# 关于  \n这是面向 1.0.0 版本的一大批变更中的首个非预发布版本。我们决定进行一次常规的 Beta 发布，因为预计不会出现破坏性更改，且主分支在过去几个月中一直运行稳定。与 0.6.x 版本相比，本次发布包含许多修复和改进。\n\n## 主要变更  \n_在 1.0.0rc1 中引入：_  \n* [重大] 由 @LeonieFreisinger 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1311 中实现，加速嵌套加法模型  \n* [重大] 由 @noxan 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1271 中完成，放弃对 Python 3.7 的支持并全面迁移到 Poetry  \n\n_在 1.0.0rc3 中引入：_  \n* [重大] 由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1356 中通过向量化 for 循环，加速 split_df 函数  \n\n_在 1.0.0rc4 中引入：_  \n* [重大] 由 @noxan 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1404 中完成，升级到 PyTorch 2  \n\n_在 1.0.0rc5 中引入：_  \n* [重大] 训练方面：由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1462 中将默认批大小调大  \n* [重大] 训练方面：由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1450 中将 1cycle 学习率策略更新为三阶段模式  \n* [重大] 训练方面：由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1461 中将 SmoothL1Loss 的 beta 参数从 1.0 调整为 0.3  \n\n_在 1.0.0rc6 中引入：_  \n* [重大] 当尝试第二次拟合模型时抛出 RuntimeError（修复 #1493），由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1531 中实现  \n* [重大] 对于小数据集减少批大小，对大多数数据集减少总训练轮数，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1533 中完成  \n\n## 自 0.6.0 以来的所有变更  \n* [重大] 由 @noxan 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1271 中完成，放弃对 Python 3.7 的支持并全面迁移到 Poetry  \n* [文档] 由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1301 中对功能指南进行全面改版  \n* [网站] 由 @noxan 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1307 中修复使用 Poetry 部署网站的问题  \n* 由 @christymctse 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1225 中提供的不确定性量化教程  \n* [次要] 由 @hxyue1 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1286 中添加静态 Plotly 后端  \n* [文档] 由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1309 中更新文档网站  \n* [次要] 由 @noxan 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1305 中将依赖项更新至最新版本  \n* [文档] 由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1310 中将可复现性加入教程 10  \n* [次要] 由 @noxan 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1316 中统一版本号处理、移除 setup.py 并迁移相关内容  \n* [重构] 由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_","2024-02-14T23:28:31",{"id":194,"version":195,"summary_zh":196,"released_at":197},324712,"1.0.0rc6","## 主要变更\n* [重大] 当尝试第二次拟合模型时，抛出 RuntimeError（修复 #1493）由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1531 中完成\n* [重大] 对于小型数据集减少批次大小，对大多数数据集减少整体训练轮数由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1533 中完成\n\n## 修复\n* [修复] 在使用不同加速器的机器上加载模型由 @McOffsky 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1509 中完成\n* [修复] 绘图参数 - 年度季节性由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1501 中完成\n\n## 文档\n* [文档] 修复 MLFlow Sphinx 测试错误由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1508 中完成\n* [文档] MLflow 教程由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1494 中完成\n* [文档] 向 MLflow_and_NeuralProphet.ipynb 添加实验由 @cargecla1 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1499 中完成\n\n## 依赖与测试\n* [DevOps] 将 jinja2 从 3.1.2 升级到 3.1.3由 @dependabot 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1506 中完成\n* [DevOps] 升级到 Pandas 2.0.0由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1498 中完成\n* [DevOps] 升级到 torch >=2.0.0由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1512 中完成\n* [DevOps] 支持 Python 3.11由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1513 中完成\n* [DevOps] 将测试升级至 Python 3.11，并移除 Windows 支持由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1515 中完成\n* [DevOps] 更新 Poetry 开发依赖由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1530 中完成\n\n## 新贡献者\n* @McOffsky 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1509 中完成了首次贡献\n* @cargecla1 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1499 中完成了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F1.0.0rc5...1.0.0rc6","2024-02-14T22:55:52",{"id":199,"version":200,"summary_zh":201,"released_at":202},324713,"1.0.0rc5","## 变更内容\n### 重大变更\n* [重大] 训练：默认批量大小增加，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1462 中提出\n* [重大] 训练：将 1cycle 学习率策略更新为 3 阶段策略，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1450 中提出\n* [重大] 训练：SmoothL1Loss 的 beta 参数从 1.0 更新为 0.3，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1461 中提出\n\n### 次要变更\n* [次要] 辅助函数支持无日期戳的等距数据，由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1432 中提出\n* [次要] 添加 Torch 加载功能，由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1473 中提出\n\n### Bug 修复\n* [修复] 减少保存模型的大小，由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1425 中提出\n* [修复] 修正节假日测试并更新版本号，由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1431 中提出\n* [修复] 为修复绘制事件组件时出现的 ValueError 错误添加 Pytest 测试，由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1434 中提出\n* [修复] 添加绘图相关的 Pytest 测试，由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1445 中提出\n\n### 文档更新\n* [文档] 移除不正确的注释，由 @thekoc 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1421 中提出\n* [文档] 修正 README.md 中的拼写错误，由 @eltociear 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1436 中提出\n* [文档] 更新 CONTRIBUTING.md 中关于 Poetry 开发安装的部分，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1357 中提出\n* [文档] 修正教程中关于数据描述的部分，将“每日”改为“每小时”，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1453 中提出\n* [文档] 正确提及 SmoothL1Loss 而非 Huber 损失，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1458 中提出\n* [文档] 修正 README.md 中的拼写错误，由 @VinayKokate22 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1472 中提出\n* [文档] 修复拼写错误，由 @Mayureshd-18 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1470 中提出\n* [文档] 更新 README.md，由 @harshhere905 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1467 中提出\n* [文档] 更新滞后回归量教程，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1463 中提出\n* [文档] 添加句号，由 @bhargavshirin 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1465 中提出\n* [文档] 在 README.md 中添加贡献者部分，由 @Kalyanimhala 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1474 中提出\n* [文档] 修复 df_utils.py 中的拼写错误，由 @eltociear 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1478 中提出\n\n### 测试改进\n* [测试] 扩展模型测试（部分在 #1464 中撤销），由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1455 中提出\n* [测试] 添加新测试并移除 3 个重复测试，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1464 中提出\n\n### DevOps 改进\n* [DevOps] 将版本号提升至 1.0.0rc5，由 @ourownstory 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1440 中提出\n* [DevOps] 更新 poetry lo","2023-12-11T19:31:34",{"id":204,"version":205,"summary_zh":206,"released_at":207},324714,"1.0.0rc4","## 变更内容\n\n### 重大变更\n* [重大] 升级到 PyTorch 2，由 @noxan 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1404 中完成\n\n### 小幅变更\n* [小幅] 内存优化的 stride 函数，由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1400 中完成\n* [小幅] 加快数据加载器速度，由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1360 中完成\n* [小幅] 使用内存效率更高的 float32 类型替代 float64 类型，由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1402 中完成\n* [小幅] 向 NeuralProphet.test 添加 verbose 选项，由 @c3-ziqin 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1407 中完成\n\n### 已修复的缺陷\n* [缺陷] 修复 PyTorch Lightning 模块中 torchmetrics 的正确定义，由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1365 中完成\n* [缺陷] 修复 ruff 警告，由 @leoniewgnr 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1408 中完成\n* [修复] 移除错误的绘图警告，由 @Maisa-bs 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1411 中完成\n* 在将模型保存到文件后，将训练器重新添加回预测器中，由 @c3-ziqin 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1416 中完成\n* [修复] Colab 上的教程崩溃问题，由 @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1418 中完成\n\n## 新贡献者\n* @c3-ziqin 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1407 中完成了首次贡献\n* @Maisa-bs 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1411 中完成了首次贡献\n* @SimonWittner 在 https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1418 中完成了首次贡献\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F1.0.0rc3...1.0.0rc4","2023-09-19T18:21:07",{"id":209,"version":210,"summary_zh":211,"released_at":212},324715,"1.0.0rc3","## What's Changed\r\n## Major\r\n* [major] Speed up split_df by vectorizing for loops by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1356\r\n\r\n## Minor\r\n* update selective forecasting tests by @ourownstory in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1312\r\n* [devops] Remove macos and windows from Github actions by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1325\r\n* [devops] Add ruff github action by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1324\r\n* [refactor] Ruff fixes by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1327\r\n* Remove type conversion for _stride_lagged_features and _stride_future… by @hxyue1 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1331\r\n* [minor] fixed prediction_frequency when quantiles are set by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1347\r\n* Update poetry dependencies by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1390\r\n* [minor] fix sum of components  by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1349\r\n\r\n## Bug fixes\r\n* [bug] Fix pyright and flake8 warnings by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1392\r\n* [bug] Trend plotting when using numpy 1.24 by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1395\r\n* (Fix) pytorch_lightning\u003C2.0.0 callbacks named ProgressBarBase instead of ProgressBar by @JSarsfield in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1344\r\n* Fixes failing tests by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1381\r\n\r\n\r\n## Docs\r\n* [docs] fix showing of plots in tutorials and feature guides by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1322\r\n* [docs] prettify application examples by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1328\r\n* [docs] Fix prophet tutorial by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1332\r\n* [docs] update prophet tutorials by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1334\r\n* [docs] Clear plotting tutorial by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1343\r\n* fix tutorials link in readme by @ShreyaKhurana in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1353\r\n* [docs] ar layers fixed in uncertainty tutorial by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1364\r\n* [docs] Small fixes in documentation by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1380\r\n* [docs] New feature guide with industry application and tool by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1369\r\n\r\n## New Contributors\r\n* @ShreyaKhurana made their first contribution in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1353\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F1.0.0rc2...1.0.0rc3","2023-08-14T20:26:08",{"id":214,"version":215,"summary_zh":216,"released_at":217},324716,"0.6.2","## What's Changed\r\n* fixed assert error by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1354\r\n* fixed component computation by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1350","2023-06-14T22:41:07",{"id":219,"version":220,"summary_zh":221,"released_at":222},324717,"1.0.0rc2","## What's Changed\r\n* [fix] Downgrade kaleido version to 0.2.1 to fix installs by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1320\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F1.0.0rc1...1.0.0rc2","2023-05-02T02:02:06",{"id":224,"version":225,"summary_zh":226,"released_at":227},324718,"1.0.0rc1","## What's Changed\r\n\r\n### Features\r\n* [major] speed up nested additive model by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1311\r\n* [major] Drop Python 3.7 support and complete poetry migration by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1271\r\n* [minor] Add static plotly backend by @hxyue1 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1286\r\n* [minor] Update dependencies to latest versions by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1305\r\n* [minor] Unify version number handling, remove setup.py and migrate content by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1316\r\n\r\n\r\n### Docs\r\n* [docs] Revamp of feature guides by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1301\r\n* [website] Fix the deployment of the website with poetry by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1307\r\n* [docs] Uncertainty quantification tutorial by @christymctse in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1225\r\n* [docs] Update documentation website by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1309\r\n* [docs] added reproducibility to tutorial 10 by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1310\r\n\r\n\r\n### Refactor\r\n* [refactor] Uncertainty streamline naming by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1262\r\n\r\n\r\n## New Contributors\r\n* @hxyue1 made their first contribution in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1286\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.6.0...1.0.0rc1","2023-05-02T01:43:01",{"id":229,"version":230,"summary_zh":231,"released_at":232},324719,"0.6.0","## What's Changed\r\n\r\n* [major] Nested Additive Model by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1216\r\n* [major] Allow user to set prediction frequency as a dict: {freq: forecast origin} by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1187\r\n* [major] lagged regressor with interaction modeling (shared NN) by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F903\r\n* [feature] Support for asymmetrical interval in conformal quantile regression by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1215\r\n* [minor] Allow lagged regressors with only unique values for global modeling by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1267\r\n* [minor] Dependency management with poetry by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1202\r\n* [minor] Allow lagged regressors with only unique values for global modeling by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1256\r\n* [minor] Repeat static trend component n_forecasts times instead of only once by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1254\r\n* [minor] add docstring examples by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1295\r\n* [minor] Address warnings by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1276\r\n* [minor] lighten forecaster step 2 - encapsulate _get_maybe_extend_periods and _maybe_extend_df by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1278\r\n* [minor] Update dependencies to latest versions by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1280\r\n\r\n## Fixes\r\n\r\n* [fix] Calculate major frequency percentage properly by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1264\r\n* [fix] Matplotlib and plotly plot showing aligned by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1266\r\n* [fix] Remove regressor config if all regressors are removed due to unique values  by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1272\r\n* [fix] Unregister plotly resampler directly after usage by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1277\r\n* [fix] Raise Exception again to avoid UnboundLocalError by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1282\r\n* [fix] Remove SmoothL1Loss from Tutorial by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1298\r\n* [fix] Raise ValueError when df contains not enough rows by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1300\r\n\r\n## Tests\r\n* [tests] Remove unneeded logs by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1207\r\n\r\n## Documentation\r\n* [docs] outdated labels in contributing.md by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1242\r\n* [docs] Create new tutorials content by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1183\r\n* [docs] add note on flexible multiplicativity from prophet by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1268\r\n* [docs] Add next steps to tutorials by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1274\r\n* [docs] Improve conformal prediction tutorial by adding marginal before coverage by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1284\r\n\r\n## Website\r\n\r\n* [website] Fix website deployment by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1273\r\n\r\n## Refactors\r\n* [refactor] Removed plotting in test_uncertainty.py. by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1246\r\n* [refactor] Changed int() to round() for q_hat_idx in uncertainty.py. by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1248\r\n* [refactor] Lighten forecaster.py by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1255\r\n* [refactor] Encapsulate _make_future_dataframe and _check_dataframe by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1279\r\n* [refactor] Encapsulate _validate_column_name by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1289\r\n* [refactor] Encapsulate _normalize by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1287\r\n* [refactor] Encapsulate _handle_missing_data and _handle_missing_data_single_id by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1288\r\n* [refactor] Encapsulate _prepare_dataframe_to_predict by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1290\r\n* [refactor] Encapsulate _convert_raw_predictions_to_raw_df by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1291\r\n* [refactor] Encapsulate _reshape_raw_predictions_to_forecst_df by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1293\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.5.3...0.6.0","2023-04-26T02:00:02",{"id":234,"version":235,"summary_zh":236,"released_at":237},324720,"0.5.4","## What's Changed\r\n\r\n### Documentation\r\n* [docs] Create new tutorials content by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1183\r\n* [docs] outdated labels in contributing.md by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1242\r\n\r\n### Refactors\r\n* [refactor] Removed plotting in test_uncertainty.py. by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1246\r\n* [refactor] Changed int() to round() for q_hat_idx in uncertainty.py. by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1248\r\n\r\n### Tests\r\n* [tests] Remove unneeded logs by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1207\r\n\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.5.3...0.5.4","2023-04-08T02:08:56",{"id":239,"version":240,"summary_zh":241,"released_at":242},324721,"0.5.3","## What's Changed\r\n\r\n### Bugfixes\r\n* [fix] Limit broken dependency versions (mainly pytorch \u003C 2.0) by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1221\r\n\r\n### Minor features\r\n* [minor] Adjust seasonality reg to modularized code by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1203\r\n\r\n### Documentation\r\n* [docs] Style fixes in tutorial notebooks when dark mode is active by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1199\r\n* [typing] Add typing for user facing functions by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1193\r\n\r\n### Under the hood changes\r\n* [refactor] Move evaluate method outside of Conformal class and rename it as conformal_evaluate by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1210\r\n* [refactor] Renamed conformal.py to uncertainty.py and conformal_evalute() to uncertainty_evaluate(). by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1213\r\n* [refactor] Modified n_forecasts in  _infer_evaluate_params_from_dataset() method in conformal.py by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1212\r\n* [refactor] Add quantile regression (QR) as an uncertainty evalution metric in uncertainty.py by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1214\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.5.2...0.5.3","2023-03-17T03:18:45",{"id":244,"version":245,"summary_zh":246,"released_at":247},324722,"0.5.2","# What's changed\r\n\r\n## [breaking] Breaking changes\r\n* Change default plotting backend to plotly-resample by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1149\r\n\r\n## [major] Major features \r\n* Trend modularization by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1078\r\n* Seasonality modularization by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1141\r\n* Future regressors modularization by @alfonsogarciadecorral in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1144\r\n\r\n\r\n## [minor] Minor changes \r\n* Allow unique values for future regressor of one time series in global modeling by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1146\r\n* Added ability to pass custom callbacks while adhering to callback con… by @JSarsfield in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1122\r\n* [minor] Add get function for parameters by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1175\r\n* Improve usability for conditional seasonality by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1116\r\n* [Feature] Enable own NN configuration for lagged regressors by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1154\r\n* [Minor] Convert timezones to UTC by default by @christymctse in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1085\r\n* Typing improvements by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1121\r\n* Typing conformal prediction by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1124\r\n* [refactor] create dict for events or regressors by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1123\r\n* [refactor] Extract holiday getter helper to single function by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1072\r\n* black edited utils.py by @ourownstory in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1160\r\n* Uncertainty: Conformal Prediction V1.2 - move conformal evaluation performance metrics inside of Conformal class in conformal.py by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1155\r\n* [refactor] incorporate typing for time_net.py by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1105\r\n* [refactor] Update python-holidays integration by @arkid15r in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1176\r\n\r\n## [fix] Bugfixes \r\n* Fix ipython version for plotly-resampler by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1119\r\n* Deactivate plotly resampler for static image export by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1114\r\n* Lightning version by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1145\r\n* Avoidance of duplicate code in regularisation by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1110\r\n\r\n## [docs] Documentation \r\n* Created narratives and simplified code for quantile regression tutorial material by @christymctse in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1091\r\n* Update CONTRIBUTING.md by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1169\r\n* [fix] default value of n_changepoints given correctly in the docs by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1128\r\n* [docs] Readme library purpose by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1140\r\n\r\n## [tests] Tests \r\n* [tests] Added info logs to regularisation tests by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1148\r\n\r\n## [devops] Github workflows \r\n* Fixed pyright error raised from unbound list by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1156\r\n\r\n# New Contributors\r\n* @JSarsfield made their first contribution in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1122\r\n* @arkid15r made their first contribution in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1176\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.5.1...0.5.2","2023-03-06T20:41:24",{"id":249,"version":250,"summary_zh":251,"released_at":252},324723,"0.5.1","## New \u002F major improvements\r\n* Introduce conditional seasonality by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1067\r\n* add plotly resampler by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1083\r\n* Uncertainty: Conformal Prediction V1.1 - extend to multiple forecast steps by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1073\r\n\r\n\r\n## Breaking changes\r\n* Progress bar in minimal mode is default instead of deactivated by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1059\r\n\r\n## List of all changes\r\n* Uncertainty: Conformal Prediction V1.1 - add Conformal class to conformal_prediction.py and rename file to conformal.py by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1074\r\n* [Fix] Fixed error in plot loss by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1064\r\n* Restict numpy version to \u003C1.24.0 by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1080\r\n* remove default settings (None for calibration df and 0.1 for alpha) by @christymctse in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1079\r\n* [Bug-fix] Enable test_plot_conformal_prediction with auto-regression on plot_latest_forecast by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1096\r\n* [refactor] incorporate typing for Conformal class and conformal_predict method by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1104\r\n* [Bug-fix] Enable holidays when using global\u002Fglocal modeling by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1100\r\n* [Bug-fix] Refactor warning about global normalization to forecaster.py by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1109\r\n\r\n## New Contributor\r\n* @christymctse made their first contribution in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1079\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fcompare\u002F0.5.0...0.5.1","2023-01-15T20:50:05",{"id":254,"version":255,"summary_zh":256,"released_at":257},324724,"0.5.0","## New \u002F major improvements\r\n* Lightning Migration by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F837\r\n* GPU support by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F961\r\n* Support plotting of only specific panels in plot_parameters and plot_components by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F884\r\n* Uncertainty: Conformal Prediction V1 with .conformal_predict() by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F802 and https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1044\r\n* Support for holidays of multiple countries by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1001\r\n* Deep model parameter interpretation by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F883\r\n* Add type annotations for main NeuralProphet class by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F981\r\n\r\n## Breaking changes\r\n\r\n* Remove support for df dict as an input by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F929\r\n* Change plotting_backend deprecation warning for implicit and explicit matplotlib use by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1006\r\n* Refactored fit() interface by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F1039\r\n* Remove residuals in forecast dataframe returned by m.predict() by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F927\r\n\r\n## List of all changes\r\n* [lightning] Remove rich progress bar by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F952\r\n* Improvement\u002Ftype annotations by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F954\r\n* Commented out test_progress_display in test_integration.py. by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F956\r\n* Added plot param of the NeuralProphet fit() and uncomment test_progress_display in test_integration.py. by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F958\r\n* [lightning] refactored Trainer args by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F959\r\n* refactored test failure by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F960\r\n* Add alias function plot_last_forecast() for plot_latest_forecast() and give deprecation warning by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F919\r\n* Mask seasonality on Sat\u002FSun when data frequency is Business days by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F808\r\n* [bug] Forecast output dataframe contains all float dtypes instead of having some objects by @Kevin-Chen0 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F978\r\n* Fix labels of yhatx and origin-x in plots by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F942\r\n* Improve github action speed for docs test by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F936\r\n* Resolved discontinuous trend when changepoints are not sorted chronologically by @leoniewgnr in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F968\r\n* Address small tech debt to use a conditional instead of try\u002Fexcept for Python 3.7 support by @kapoor1992 in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F934\r\n* [dev-ops] Collapsible metrics reports by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F926\r\n* Cleanup document files by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F975\r\n* Rename plot_forecast.py and plot_model_parameters.py to xxx_matplotli… by @SaumyaBhushan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F971\r\n* [naming] Renamed tutorial notebooks by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F921\r\n* Labels in CONTRIBUTING.md by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F946\r\n* [enhancement] add pytests for matplotlib  by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F914\r\n* [devops] Improve codecov threshold by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F980\r\n* [devops] Upgrade GitHub action versions by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F979\r\n* change deprecation message to switch plotting backend to plotly by @LeonieFreisinger in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F984\r\n* [feature] Plot components of averaged df (and quants) if multiple IDs, but no df_name specified in plot_components() by @judussoari in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F888\r\n* [devops] Code coverage threshold fix by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F990\r\n* Time in model training benchmark by @karl-richter in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F995\r\n* Unified the naming of variables - num and nb by @SaumyaBhushan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F994\r\n* [tests] Add tests for configure by @noxan in https:\u002F\u002Fgithub.com\u002Fourownstory\u002Fneural_prophet\u002Fpull\u002F983\r\n* [cleanup]","2022-12-08T00:04:31"]