[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-Nixtla--nixtla":3,"tool-Nixtla--nixtla":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":75,"owner_avatar_url":76,"owner_bio":77,"owner_company":78,"owner_location":78,"owner_email":79,"owner_twitter":80,"owner_website":81,"owner_url":82,"languages":83,"stars":110,"forks":111,"last_commit_at":112,"license":113,"difficulty_score":23,"env_os":114,"env_gpu":114,"env_ram":114,"env_deps":115,"category_tags":120,"github_topics":121,"view_count":10,"oss_zip_url":78,"oss_zip_packed_at":78,"status":16,"created_at":137,"updated_at":138,"faqs":139,"releases":168},772,"Nixtla\u002Fnixtla","nixtla","TimeGPT-1: production ready pre-trained Time Series Foundation Model  for forecasting and anomaly detection. Generative pretrained transformer for time series trained on over 100B data points. It's capable of accurately predicting various domains such as retail, electricity, finance, and IoT with just a few lines of code 🚀.","nixtla 是一款专注于时间序列分析的开源 AI 工具，其核心 TimeGPT-1 是首个面向预测与异常检测的基础模型。它解决了传统时序建模依赖人工特征工程、开发周期长且难以跨领域复用的痛点。凭借在超 1000 亿数据点上训练的生成式预训练 Transformer，nixtla 能够以极低的代码成本，精准处理零售、电力、金融及物联网等多领域的预测需求。\n\n对于开发者而言，nixtla 极大降低了使用门槛，仅需寥寥数行 Python 代码即可完成从数据加载到未来趋势预测的全过程，同时也支持异常检测功能。工具还提供多语言 API 支持，并具备独特的 Snowflake 部署能力，允许在不移动数据的前提下直接在云数据仓库中运行预测服务。无论是进行数据分析的研究人员，还是寻求快速落地的企业开发者，都能借助 nixtla 轻松实现高效的时间序列智能应用。","# Nixtla   [![Tweet](https:\u002F\u002Fimg.shields.io\u002Ftwitter\u002Furl\u002Fhttp\u002Fshields.io.svg?style=social)](https:\u002F\u002Ftwitter.com\u002Fintent\u002Ftweet?text=Statistical%20Forecasting%20Algorithms%20by%20Nixtla%20&url=https:\u002F\u002Fgithub.com\u002FNixtla\u002Fneuralforecast&via=nixtlainc&hashtags=StatisticalModels,TimeSeries,Forecasting)  [![Slack](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FSlack-4A154B?&logo=slack&logoColor=white)](https:\u002F\u002Fjoin.slack.com\u002Ft\u002Fnixtlacommunity\u002Fshared_invite\u002Fzt-1pmhan9j5-F54XR20edHk0UtYAPcW4KQ)\n\n\u003Cdiv align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_f5bda935f14f.png\"\u002F>\n\u003Ch1 align=\"center\">TimeGPT-1 \u003C\u002Fh1>\n\u003Ch3 align=\"center\">The first foundation model for forecasting and anomaly detection\u003C\u002Fh3>\n\n[![CI](https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Factions\u002Fworkflows\u002Fci.yaml\u002Fbadge.svg?branch=main)](https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Factions\u002Fworkflows\u002Fci.yaml)\n[![PyPi](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fnixtla?color=blue)](https:\u002F\u002Fpypi.org\u002Fproject\u002Fnixtla\u002F)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache_2.0-blue.svg)](https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fblob\u002Fmain\u002FLICENSE)\n[![docs](https:\u002F\u002Fimg.shields.io\u002Fwebsite-up-down-green-red\u002Fhttp\u002Fdocs.nixtla.io\u002F.svg?label=docs)](https:\u002F\u002Fdocs.nixtla.io)\n[![Downloads](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_34c50e933400.png)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fnixtla)\n[![Downloads](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_34c50e933400.png\u002Fmonth)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fnixtla)\n[![Downloads](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_34c50e933400.png\u002Fweek)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fnixtla)\n\n**TimeGPT** is a production ready, generative pretrained transformer for time series. It's capable of accurately predicting various domains such as retail, electricity, finance, and IoT with just a few lines of code 🚀.\n\n\u003C\u002Fdiv>\n\n## 🚀 Quick Start\n\nhttps:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fassets\u002F4086186\u002F163ad9e6-7a16-44e1-b2e9-dab8a0b7b6b6\n\n### Install nixtla's SDK\n\n```python\npip install nixtla>=0.7.0\n```\n\n### Import libraries and load data\n\n```python\nimport pandas as pd\nfrom nixtla import NixtlaClient\n```\n\n### Forecast using TimeGPT in 3 easy steps\n\n```python\n# Get your API Key at https:\u002F\u002Fnixtla.io\u002Ffree-trial?utm_source=nixtla.io&utm_campaign=\u002Fdocs\u002Freadme\n\n# 1. Instantiate the NixtlaClient\nnixtla_client = NixtlaClient(api_key = 'YOUR API KEY HERE')\n\n# 2. Read historic electricity demand data\ndf = pd.read_csv('https:\u002F\u002Fraw.githubusercontent.com\u002FNixtla\u002Ftransfer-learning-time-series\u002Fmain\u002Fdatasets\u002Felectricity-short.csv')\n\n# 3. Forecast the next 24 hours\nfcst_df = nixtla_client.forecast(df, h=24, level=[80, 90])\n\n# 4. Plot your results (optional)\nnixtla_client.plot(df, fcst_df, level=[80, 90])\n\n```\n\n![Forecast Results](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_3a6a6609e18c.png)\n\n### Anomaly detection using TimeGPT in 3 easy steps\n\n```python\n# Get your API Key at https:\u002F\u002Fnixtla.io\u002Ffree-trial?utm_source=nixtla.io&utm_campaign=\u002Fdocs\u002Freadme\n\n# 1. Instantiate the NixtlaClient\nnixtla_client = NixtlaClient(api_key = 'YOUR API KEY HERE')\n\n# 2. Read Data # Wikipedia visits of NFL Star (\ndf = pd.read_csv('https:\u002F\u002Fdatasets-nixtla.s3.amazonaws.com\u002Fpeyton-manning.csv')\n\n\n# 3. Detect Anomalies\nanomalies_df = nixtla_client.detect_anomalies(df, time_col='timestamp', target_col='value', freq='D')\n\n# 4. Plot your results (optional)\nnixtla_client.plot(df, anomalies_df,time_col='timestamp', target_col='value')\n```\n\n![AnomalyDetection](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_7731fdbfb68f.png)\n\n## 🤓 API support for other languages\n\nExplore our [API Reference](https:\u002F\u002Fdocs.nixtla.io) to discover how to leverage TimeGPT across various programming languages including JavaScript, Go, and more.\n\n## ❄️ Snowflake Deployment\n\nRun TimeGPT directly within your Snowflake environment. The deployment script creates stored procedures and UDTFs that enable forecasting and anomaly detection on your Snowflake data without moving it outside your infrastructure.\n\n```bash\npip install nixtla[snowflake]\npython -m nixtla.scripts.snowflake_install_nixtla\n```\n\nThe script will guide you through setting up external access integrations, configuring your API key, and deploying the forecasting components to your specified database and schema.\n\n## 🔥 Features and Capabilities\n\n- **Zero-shot Inference**: TimeGPT can generate forecasts and detect anomalies straight out of the box, requiring no prior training data. This allows for immediate deployment and quick insights from any time series data.\n\n- **Fine-tuning**: Enhance TimeGPT's capabilities by fine-tuning the model on your specific datasets, enabling the model to adapt to the nuances of your unique time series data and improving performance on tailored tasks.\n\n- **API Access**: Integrate TimeGPT seamlessly into your applications via our robust API. Upcoming support for Azure Studio will provide even more flexible integration options. Alternatively, deploy TimeGPT on your own infrastructure to maintain full control over your data and workflows.\n\n- **Add Exogenous Variables**: Incorporate additional variables that might influence your predictions to enhance forecast accuracy. (E.g. Special Dates, events or prices)\n\n- **Multiple Series Forecasting**: Simultaneously forecast multiple time series data, optimizing workflows and resources.\n\n- **Custom Loss Function**: Tailor the fine-tuning process with a custom loss function to meet specific performance metrics.\n\n- **Cross Validation**: Implement out of the box cross-validation techniques to ensure model robustness and generalizability.\n\n- **Prediction Intervals**: Provide intervals in your predictions to quantify uncertainty effectively.\n\n- **Irregular Timestamps**: Handle data with irregular timestamps, accommodating non-uniform interval series without preprocessing.\n\n## 📚 Documentation with examples and use cases\n\nDive into our [comprehensive documentation](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart) to discover examples and practical use cases for TimeGPT. Our documentation covers a wide range of topics, including:\n\n- **Getting Started**: Begin with our user-friendly [Quickstart Guide](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart) and learn how to [set up your API key](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-setting_up_your_api_key) effortlessly.\n\n- **Advanced Techniques**: Master advanced forecasting methods and learn how to enhance model accuracy with our tutorials on [anomaly detection](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-anomaly_detection), fine-tuning models using specific loss functions, and scaling computations across distributed frameworks such as [Spark, Dask, and Ray](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-computing_at_scale).\n\n- **Specialized Topics**: Explore specialized topics like [handling exogenous variables](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-holidays_and_special_dates), model validation through [cross-validation](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-cross_validation), and strategies for [forecasting under uncertainty](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-uncertainty_quantification).\n\n- **Real World Applications**: Uncover how TimeGPT is applied in real-world scenarios through case studies on [forecasting web traffic](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fuse-cases-forecasting_web_traffic) and [predicting Bitcoin prices](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fuse-cases\u002Fbitcoin_price_prediction).\n\n## 🗞️ TimeGPT1 Revolutionizing Forecasting and Anomaly Detection\n\nTime series data is pivotal across various sectors, including finance, healthcare, meteorology, and social sciences. Whether it's monitoring ocean tides or tracking the Dow Jones's daily closing values, time series data is crucial for forecasting and decision-making.\n\nTraditional analysis methods such as ARIMA, ETS, MSTL, Theta, CES, machine learning models like XGBoost and LightGBM, and deep learning approaches have been standard tools for analysts. However, TimeGPT introduces a paradigm shift with its standout performance, efficiency, and simplicity. Thanks to its zero-shot inference capability, TimeGPT streamlines the analytical process, making it accessible even to users with minimal coding experience.\n\nTimeGPT is user-friendly and low-code, enabling users to upload their time series data and either generate forecasts or detect anomalies with just a single line of code. As the only foundation model for time series analysis out of the box, TimeGPT can be integrated via our public APIs, through Azure Studio (coming soon), or deployed on your own infrastructure.\n\n## ⚙️ TimeGPT's Architecture\n\nSelf-attention, the revolutionary concept introduced by the paper “Attention is all you need“, is the basis of the this foundational model. The TimeGPT model is not based on any existing large language model(LLMs). It is independently trained on vast timeseries dataset as a large transformer model and is designed so as to minimize the forecasting error.\n\nThe architecture consists of an encoder-decoder structure with\nmultiple layers, each with residual connections and layer normalization. Finally, a linear layer maps the decoder’s output to the forecasting window dimension. The general intuition is that attentionbased mechanisms are able to capture the diversity of past events and correctly extrapolate potential\nfuture distributions.\n\n![Arquitecture](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_3179ad7e139d.png)\n\nTimeGPT was trained on, to our knowledge, the largest collection of publicly available time series,\ncollectively encompassing over 100 billion data points. This training set incorporates time series\nfrom a broad array of domains, including finance, economics, demographics, healthcare, weather,\nIoT sensor data, energy, web traffic, sales, transport, and banking. Due to this diverse set of domains,\nthe training dataset contains time series with a wide range of characteristics\n\n---\n\n## ⚡️ Zero-shot Results\n\n### Accuracy\n\nTimeGPT has been tested for its zero-shot inference capabilities on more than 300K unique series, which involve using the model without additional fine-tuning on the test dataset. TimeGPT outperforms a comprehensive range of well-established statistical and cutting-edge deep learning models, consistently ranking among the top three performers across various frequencies.\n\n### Ease of use\n\nTimeGPT also excels by offering simple and rapid predictions using a pre-trained model. This stands in stark contrast to other models that typically require an extensive training and prediction pipeline.\n\n![Results](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_d32394dc9332.jpg)\n\n### Efficiency and Speed\n\nFor zero-shot inference, our internal tests recorded an average GPU inference speed of 0.6 milliseconds per series for TimeGPT, which nearly mirrors that of the simple Seasonal Naive.\n\n## 📝 How to cite?\n\nIf you find TimeGPT useful for your research, please consider citing the associated [paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.03589):\n\n```\n@misc{garza2023timegpt1,\n      title={TimeGPT-1},\n      author={Azul Garza and Max Mergenthaler-Canseco},\n      year={2023},\n      eprint={2310.03589},\n      archivePrefix={arXiv},\n      primaryClass={cs.LG}\n}\n```\n\n## 🎉 Features and Mentions\n\nTimeGPT has been featured in many publications and has been recognized for its innovative approach to time series forecasting. Here are some of the features and mentions:\n\n- [TimeGPT Revolutionizing Time Series Forecasting](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2024\u002F02\u002Ftimegpt-revolutionizing-time-series-forecasting\u002F)\n- [TimeGPT: The First Foundation Model for Time Series Forecasting](https:\u002F\u002Ftowardsdatascience.com\u002Ftimegpt-the-first-foundation-model-for-time-series-forecasting-bf0a75e63b3a)\n- [TimeGPT: Revolutionising Time Series Forecasting with Generative Models](https:\u002F\u002Fmedium.com\u002F@22meera99\u002Ftimegpt-revolutionising-time-series-forecasting-with-generative-models-86be6c09fa51)\n- [TimeGPT on Turing Post](https:\u002F\u002Fwww.turingpost.com\u002Fp\u002Ftimegpt)\n- [TimeGPT Presentation at AWS Events](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=5pYkT0rTCfE&ab_channel=AWSEvents)\n- [TimeGPT: Machine Learning for Time Series Made Accessible - Podcast](https:\u002F\u002Fpodcasts.apple.com\u002Fbg\u002Fpodcast\u002Ftimegpt-machine-learning-for-time-series-made-accessible\u002Fid1487704458?i=1000638551991)\n- [TimeGPT on The Data Exchange](https:\u002F\u002Fthedataexchange.media\u002Ftimegpt\u002F)\n- [How TimeGPT Transforms Predictive Analytics with AI](https:\u002F\u002Fhackernoon.com\u002Fhow-timegpt-transforms-predictive-analytics-with-ai)\n- [TimeGPT: The First Foundation Model - AI Horizon Forecast](https:\u002F\u002Faihorizonforecast.substack.com\u002Fp\u002Ftimegpt-the-first-foundation-model)\n\n## 🔖 License\n\nTimeGPT is closed source. However, this SDK is open source and available under the Apache 2.0 License. Feel free to contribute (check out the [Contributing](https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fblob\u002Fmain\u002FCONTRIBUTING.md) guide for more details).\n\n## 🏷️ Attribution\n\nNixtlaClient may be used to access services powered by technology from Google, Amazon, IBM, Datadog, and NXAI. All trademarks are the property of their respective owners.\n\n## 📞 Get in touch\n\nFor any questions or feedback, please feel free to reach out to us at ops [at] nixtla.io.\n","# Nixtla   [![推文](https:\u002F\u002Fimg.shields.io\u002Ftwitter\u002Furl\u002Fhttp\u002Fshields.io.svg?style=social)](https:\u002F\u002Ftwitter.com\u002Fintent\u002Ftweet?text=Statistical%20Forecasting%20Algorithms%20by%20Nixtla%20&url=https:\u002F\u002Fgithub.com\u002FNixtla\u002Fneuralforecast&via=nixtlainc&hashtags=StatisticalModels,TimeSeries,Forecasting)  [![Slack](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FSlack-4A154B?&logo=slack&logoColor=white)](https:\u002F\u002Fjoin.slack.com\u002Ft\u002Fnixtlacommunity\u002Fshared_invite\u002Fzt-1pmhan9j5-F54XR20edHk0UtYAPcW4KQ)\n\n\u003Cdiv align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_f5bda935f14f.png\"\u002F>\n\u003Ch1 align=\"center\">TimeGPT-1 \u003C\u002Fh1>\n\u003Ch3 align=\"center\">首个用于预测和异常检测的基础模型\u003C\u002Fh3>\n\n[![CI](https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Factions\u002Fworkflows\u002Fci.yaml\u002Fbadge.svg?branch=main)](https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Factions\u002Fworkflows\u002Fci.yaml)\n[![PyPi](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fnixtla?color=blue)](https:\u002F\u002Fpypi.org\u002Fproject\u002Fnixtla\u002F)\n[![许可证](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache_2.0-blue.svg)](https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fblob\u002Fmain\u002FLICENSE)\n[![文档](https:\u002F\u002Fimg.shields.io\u002Fwebsite-up-down-green-red\u002Fhttp\u002Fdocs.nixtla.io\u002F.svg?label=docs)](https:\u002F\u002Fdocs.nixtla.io)\n[![下载量](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_34c50e933400.png)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fnixtla)\n[![下载量](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_34c50e933400.png\u002Fmonth)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fnixtla)\n[![下载量](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_34c50e933400.png\u002Fweek)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fnixtla)\n\n**TimeGPT** 是一款面向时间序列的生产就绪型生成式预训练 Transformer（变换器）。它仅需几行代码即可准确预测零售、电力、金融和物联网（IoT）等多个领域 🚀。\n\n\u003C\u002Fdiv>\n\n## 🚀 快速开始\n\nhttps:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fassets\u002F4086186\u002F163ad9e6-7a16-44e1-b2e9-dab8a0b7b6b6\n\n### 安装 nixtla 的 SDK\n\n```python\npip install nixtla>=0.7.0\n```\n\n### 导入库并加载数据\n\n```python\nimport pandas as pd\nfrom nixtla import NixtlaClient\n```\n\n### 使用 TimeGPT 进行预测的 3 个简单步骤\n\n```python\n# Get your API Key at https:\u002F\u002Fnixtla.io\u002Ffree-trial?utm_source=nixtla.io&utm_campaign=\u002Fdocs\u002Freadme\n\n# 1. Instantiate the NixtlaClient\nnixtla_client = NixtlaClient(api_key = 'YOUR API KEY HERE')\n\n# 2. Read historic electricity demand data\ndf = pd.read_csv('https:\u002F\u002Fraw.githubusercontent.com\u002FNixtla\u002Ftransfer-learning-time-series\u002Fmain\u002Fdatasets\u002Felectricity-short.csv')\n\n# 3. Forecast the next 24 hours\nfcst_df = nixtla_client.forecast(df, h=24, level=[80, 90])\n\n# 4. Plot your results (optional)\nnixtla_client.plot(df, fcst_df, level=[80, 90])\n\n```\n\n![Forecast Results](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_3a6a6609e18c.png)\n\n### 使用 TimeGPT 进行异常检测的 3 个简单步骤\n\n```python\n# Get your API Key at https:\u002F\u002Fnixtla.io\u002Ffree-trial?utm_source=nixtla.io&utm_campaign=\u002Fdocs\u002Freadme\n\n# 1. Instantiate the NixtlaClient\nnixtla_client = NixtlaClient(api_key = 'YOUR API KEY HERE')\n\n# 2. Read Data # Wikipedia visits of NFL Star (\ndf = pd.read_csv('https:\u002F\u002Fdatasets-nixtla.s3.amazonaws.com\u002Fpeyton-manning.csv')\n\n\n# 3. Detect Anomalies\nanomalies_df = nixtla_client.detect_anomalies(df, time_col='timestamp', target_col='value', freq='D')\n\n# 4. Plot your results (optional)\nnixtla_client.plot(df, anomalies_df,time_col='timestamp', target_col='value')\n```\n\n![AnomalyDetection](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_7731fdbfb68f.png)\n\n## 🤓 其他语言的 API 支持\n\n探索我们的 [API 参考文档](https:\u002F\u002Fdocs.nixtla.io)，了解如何在包括 JavaScript、Go 等在内的各种编程语言中利用 TimeGPT。\n\n## ❄️ Snowflake 部署\n\n直接在您的 Snowflake 环境中运行 TimeGPT。部署脚本会创建存储过程和 UDTF（用户定义表函数），使您能够在不移出基础设施的情况下对 Snowflake 数据进行预测和异常检测。\n\n```bash\npip install nixtla[snowflake]\npython -m nixtla.scripts.snowflake_install_nixtla\n```\n\n该脚本将指导您设置外部访问集成、配置 API 密钥，并将预测组件部署到您指定的数据库和模式中。\n\n## 🔥 功能与能力\n\n- **零样本推理（Zero-shot Inference）**：TimeGPT 开箱即用即可生成预测和检测异常，无需先前的训练数据。这使得任何时间序列数据都能立即部署并获得快速洞察。\n\n- **微调（Fine-tuning）**：通过在特定数据集上微调模型来增强 TimeGPT 的能力，使模型能够适应您独特时间序列数据的细微差别，并在定制任务上提高性能。\n\n- **API 访问**：通过我们强大的 API 将 TimeGPT 无缝集成到您的应用程序中。即将推出的 Azure Studio 支持将提供更灵活的集成选项。或者，在您自己的基础设施上部署 TimeGPT，以保持对数据和流程的完全控制。\n\n- **添加外生变量（Exogenous Variables）**：纳入可能影响您预测的额外变量以提高预测准确性。（例如：特殊日期、事件或价格）\n\n- **多系列预测**：同时预测多个时间序列数据，优化工作流程和资源。\n\n- **自定义损失函数（Loss Function）**：使用自定义损失函数定制微调过程，以满足特定的性能指标。\n\n- **交叉验证（Cross Validation）**：实施开箱即用的交叉验证技术，以确保模型的鲁棒性和泛化性。\n\n- **预测区间**：在预测中提供区间，以有效量化不确定性。\n\n- **不规则时间戳（Timestamps）**：处理具有不规则时间戳的数据，无需预处理即可适应非均匀间隔序列。\n\n## 📚 带有示例和用例的文档\n\n深入我们的 [综合文档](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart)，探索 TimeGPT 的示例和实际用例。我们的文档涵盖广泛的主题，包括：\n\n- **入门**：从我们用户友好的 [快速入门指南](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart) 开始，轻松学习如何 [设置您的 API 密钥 (API key)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-setting_up_your_api_key)。\n\n- **高级技巧**：掌握高级预测方法，并通过我们的教程学习如何通过 [异常检测 (anomaly detection)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-anomaly_detection)、使用特定 [损失函数 (loss functions)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-anomaly_detection) 微调模型以及在分布式框架（如 [Spark, Dask 和 Ray](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-computing_at_scale)）上扩展计算来提升模型准确性。\n\n- **专业主题**：探索专业主题，例如 [处理外生变量 (exogenous variables)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-holidays_and_special_dates)、通过 [交叉验证 (cross-validation)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-cross_validation) 进行模型验证，以及 [不确定性下的预测 (forecasting under uncertainty)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Ftutorials-uncertainty_quantification) 策略。\n\n- **现实世界应用**：通过 [预测网络流量](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fuse-cases-forecasting_web_traffic) 和 [预测比特币价格](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fuse-cases\u002Fbitcoin_price_prediction) 的案例研究，揭示 TimeGPT 如何在现实场景中应用。\n\n## 🗞️ TimeGPT1 革新预测与异常检测\n\n时间序列数据 (time series data) 在金融、医疗、气象和社会科学等多个领域都至关重要。无论是监测海洋潮汐还是追踪道琼斯指数的每日收盘值，时间序列数据对于预测和决策都至关重要。\n\n传统的分析方法，如 ARIMA、ETS、MSTL、Theta、CES，机器学习模型如 XGBoost 和 LightGBM，以及 [深度学习 (deep learning)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart) 方法一直是分析师的标准工具。然而，TimeGPT 凭借其卓越的性能、效率和简洁性引入了范式转变。得益于其 [零样本推理能力 (zero-shot inference capability)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart)，TimeGPT 简化了分析流程，即使编码经验很少的用户也能轻松上手。\n\nTimeGPT 用户友好且 [低代码 (low-code)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart)，允许用户上传时间序列数据，仅用一行代码即可生成预测或检测异常。作为开箱即用的唯一 [时间序列分析基础模型 (foundation model)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart)，TimeGPT 可以通过我们的 [公共 API (public APIs)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart) 集成，或通过 Azure Studio（即将推出），或部署在您自己的基础设施上。\n\n## ⚙️ TimeGPT 的架构\n\n[自注意力机制 (Self-attention)](https:\u002F\u002Farxiv.org\u002Fabs\u002F1706.03762)，由论文《Attention is all you need》引入的革命性概念，是此基础模型的基础。TimeGPT 模型不基于任何现有的 [大型语言模型 (LLMs)](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FLarge_language_model)。它作为一个大型 [Transformer 模型](https:\u002F\u002Fpaperswithcode.com\u002Ftask\u002Ftime-series-forecasting) 在庞大的时间序列数据集上独立训练，旨在最小化预测误差。\n\n该架构由具有多层结构的 [编码器 - 解码器 (encoder-decoder)](https:\u002F\u002Farxiv.org\u002Fabs\u002F1409.3215) 组成，每层都有 [残差连接 (residual connections)](https:\u002F\u002Farxiv.org\u002Fabs\u002F1512.03385) 和 [层归一化 (layer normalization)](https:\u002F\u002Farxiv.org\u002Fabs\u002F1607.06450)。最后，一个 [线性层 (linear layer)](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FLinear_layer) 将解码器的输出映射到预测窗口维度。一般的直觉是，基于注意力的机制能够捕捉过去事件的多样性，并正确推断潜在的未来分布。\n\n![Arquitecture](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_3179ad7e139d.png)\n\n据我们所知，TimeGPT 是在最大的公开可用时间序列集合上训练的，总共包含超过 1000 亿个数据点。该训练集包含了来自广泛领域的时序数据，包括金融、经济、人口统计、医疗、天气、[物联网 (IoT)](https:\u002F\u002Fwww.ibm.com\u002Fthink\u002Ftopics\u002Finternet-of-things) 传感器数据、能源、网络流量、销售、运输和银行。由于这些多样化的领域，训练数据集包含具有各种特征的时间序列。\n\n---\n\n## ⚡️ 零样本结果\n\n### 准确性\n\nTimeGPT 已在超过 30 万个独特序列上测试了其 [零样本推理 (zero-shot inference)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart) 能力，这涉及在不针对测试数据集进行额外 [微调 (fine-tuning)](https:\u002F\u002Fdocs.nixtla.io\u002Fdocs\u002Fgetting-started-timegpt_quickstart) 的情况下使用该模型。TimeGPT 优于一系列成熟的统计模型和前沿深度学习模型，在各种频率下始终排名前三。\n\n### 易用性\n\nTimeGPT 通过使用预训练模型提供简单快速的预测，这也表现出色。这与通常需要广泛训练和预测管道的其他模型形成鲜明对比。\n\n![Results](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_readme_d32394dc9332.jpg)\n\n### 效率与速度\n\n对于零样本推理，我们的内部测试记录显示 TimeGPT 的平均 [GPU](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FGraphics_processing_unit) 推理速度为每个序列 0.6 毫秒，几乎与简单的 [季节性朴素法 (Seasonal Naive)](https:\u002F\u002Fotexts.com\u002Ffpp3\u002Fnaive.html) 持平。\n\n## 📝 如何引用？\n\n如果您发现 TimeGPT 对您的研究有用，请考虑引用相关的 [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.03589)：\n\n```\n@misc{garza2023timegpt1,\n      title={TimeGPT-1},\n      author={Azul Garza and Max Mergenthaler-Canseco},\n      year={2023},\n      eprint={2310.03589},\n      archivePrefix={arXiv},\n      primaryClass={cs.LG}\n}\n```\n\n## 🎉 功能与提及\n\nTimeGPT 已出现在许多出版物中，并因其对时间序列预测的创新方法而受到认可。以下是一些功能和提及：\n\n- [TimeGPT 革新时间序列预测](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2024\u002F02\u002Ftimegpt-revolutionizing-time-series-forecasting\u002F)\n- [TimeGPT：首个时间序列预测基础模型](https:\u002F\u002Ftowardsdatascience.com\u002Ftimegpt-the-first-foundation-model-for-time-series-forecasting-bf0a75e63b3a)\n- [TimeGPT：利用生成模型革新时间序列预测](https:\u002F\u002Fmedium.com\u002F@22meera99\u002Ftimegpt-revolutionising-time-series-forecasting-with-generative-models-86be6c09fa51)\n- [Turing Post 上的 TimeGPT](https:\u002F\u002Fwww.turingpost.com\u002Fp\u002Ftimegpt)\n- [AWS 活动上的 TimeGPT 演示](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=5pYkT0rTCfE&ab_channel=AWSEvents)\n- [TimeGPT：让时间序列机器学习触手可及 - 播客](https:\u002F\u002Fpodcasts.apple.com\u002Fbg\u002Fpodcast\u002Ftimegpt-machine-learning-for-time-series-made-accessible\u002Fid1487704458?i=1000638551991)\n- [The Data Exchange 上的 TimeGPT](https:\u002F\u002Fthedataexchange.media\u002Ftimegpt\u002F)\n- [TimeGPT 如何利用 AI 变革预测分析](https:\u002F\u002Fhackernoon.com\u002Fhow-timegpt-transforms-predictive-analytics-with-ai)\n- [TimeGPT：首个基础模型 - AI Horizon Forecast](https:\u002F\u002Faihorizonforecast.substack.com\u002Fp\u002Ftimegpt-the-first-foundation-model)\n\n## 🔖 许可证\n\nTimeGPT 是闭源的。但是，此 [SDK (软件开发工具包)](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FSoftware_development_kit) 是开源的，并根据 [Apache 2.0 许可证](https:\u002F\u002Fwww.apache.org\u002Flicenses\u002FLICENSE-2.0) 提供。欢迎贡献（查看 [贡献](https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fblob\u002Fmain\u002FCONTRIBUTING.md) 指南了解更多详情）。\n\n## 🏷️ 归属声明\n\nNixtlaClient 可用于访问由 Google、Amazon、IBM、Datadog 和 NXAI 技术提供支持的服务。所有商标均归其各自所有者所有。\n\n## 📞 联系我们\n\n如有任何问题或反馈，欢迎随时通过 ops [at] nixtla.io 与我们联系。","# Nixtla (TimeGPT-1) 快速上手指南\n\nNixtla 提供的 **TimeGPT-1** 是首个面向时间序列预测和异常检测的生成式基础模型。它无需训练即可直接进行推理，适用于零售、电力、金融等多个领域。\n\n## 环境准备\n\n*   **Python 版本**: 建议 Python 3.8 及以上\n*   **核心依赖**: `pandas`\n*   **网络要求**: 需能访问外部 API 服务（调用 TimeGPT 接口）\n\n## 安装步骤\n\n使用 pip 安装最新版本的 SDK：\n\n```bash\npip install nixtla>=0.7.0\n```\n\n> 💡 **提示**: 国内用户若下载较慢，可考虑配置国内镜像源（如清华源）：\n> ```bash\n> pip install nixtla>=0.7.0 -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n> ```\n\n## 基本使用\n\n在使用前，请先前往 [Nixtla 官网](https:\u002F\u002Fnixtla.io\u002Ffree-trial) 注册并获取免费的 **API Key**。\n\n### 1. 时间序列预测\n\n只需 3 步即可完成预测：实例化客户端、加载数据、调用预测接口。\n\n```python\nimport pandas as pd\nfrom nixtla import NixtlaClient\n\n# 1. 初始化客户端 (替换为你的 API KEY)\nnixtla_client = NixtlaClient(api_key='YOUR_API_KEY_HERE')\n\n# 2. 读取历史数据 (示例：电力需求数据)\ndf = pd.read_csv('https:\u002F\u002Fraw.githubusercontent.com\u002FNixtla\u002Ftransfer-learning-time-series\u002Fmain\u002Fdatasets\u002Felectricity-short.csv')\n\n# 3. 预测未来 24 小时\nfcst_df = nixtla_client.forecast(df, h=24, level=[80, 90])\n\n# 4. 可视化结果 (可选)\nnixtla_client.plot(df, fcst_df, level=[80, 90])\n```\n\n### 2. 异常检测\n\n同样支持零样本异常检测，识别数据中的异常点。\n\n```python\nfrom nixtla import NixtlaClient\n\n# 1. 初始化客户端\nnixtla_client = NixtlaClient(api_key='YOUR_API_KEY_HERE')\n\n# 2. 读取数据 (示例：Peyton Manning 的维基百科访问量)\ndf = pd.read_csv('https:\u002F\u002Fdatasets-nixtla.s3.amazonaws.com\u002Fpeyton-manning.csv')\n\n# 3. 检测异常\nanomalies_df = nixtla_client.detect_anomalies(\n    df, \n    time_col='timestamp', \n    target_col='value', \n    freq='D'\n)\n\n# 4. 可视化结果 (可选)\nnixtla_client.plot(df, anomalies_df, time_col='timestamp', target_col='value')\n```\n\n更多高级功能（如微调、多系列预测、Snowflake 部署等）请参考 [官方文档](https:\u002F\u002Fdocs.nixtla.io)。","某区域能源管理团队的分析师负责监控辖区内多个商业建筑的用电趋势，目标是提前预判负荷波动以降低电费支出并保障电网稳定。\n\n### 没有 nixtla 时\n- 每个建筑需单独训练统计模型，耗时数周且难以维护多套算法\n- 传统 ARIMA 等方法无法捕捉复杂非线性关系，预测误差常超 20%\n- 异常用电检测依赖人工设定阈值，漏报率高且响应严重滞后\n- 数据清洗和特征工程占据大部分开发时间，模型迭代极其缓慢\n\n### 使用 nixtla 后\n- 调用 NixtlaClient 仅需几行代码即可生成高精度预测结果，大幅降低门槛\n- 基于百亿级数据预训练的 TimeGPT 自动适应不同建筑用电模式，泛化能力强\n- 内置异常检测功能实时识别突发电力波动，无需额外开发即可发现故障\n- 支持直接对接 Snowflake 等现有数据仓库，无缝集成到生产环境无需迁移数据\n\n通过引入 nixtla，团队将预测部署周期从数周缩短至小时级，显著提升了能源调度的准确性与运维效率。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNixtla_nixtla_3a6a6609.png","Nixtla","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FNixtla_3842e028.png","Open Source Time Series Ecosystem",null,"oss@nixtla.io","nixtlainc","www.nixtla.io","https:\u002F\u002Fgithub.com\u002FNixtla",[84,88,92,96,100,104,107],{"name":85,"color":86,"percentage":87},"Jupyter Notebook","#DA5B0B",93.6,{"name":89,"color":90,"percentage":91},"MDX","#fcb32c",3.3,{"name":93,"color":94,"percentage":95},"Python","#3572A5",2.9,{"name":97,"color":98,"percentage":99},"CSS","#663399",0.1,{"name":101,"color":102,"percentage":103},"Makefile","#427819",0,{"name":105,"color":106,"percentage":103},"R","#198CE7",{"name":108,"color":109,"percentage":103},"Shell","#89e051",3820,320,"2026-04-05T02:38:17","NOASSERTION","未说明",{"notes":116,"python":114,"dependencies":117},"核心功能基于 API 调用，需申请 API Key 方可使用；支持 Snowflake 环境部署；本地部署或微调的具体硬件需求（如 GPU 型号、显存）在文档中未明确列出，仅提及有 GPU 推理速度测试数据；支持分布式计算框架如 Spark、Dask、Ray 进行扩展。",[118,119],"nixtla>=0.7.0","pandas",[15,13,26],[122,123,124,125,126,127,128,129,130,131,132,133,134,135,136],"time-series","time-series-forecasting","deep-learning","forecasting","gpt","generative-ai-time-series","timegpt","anomaly-detection","artificial-intelligence","gpts","llm","foundation-models","llms","agentic-ai","agent","2026-03-27T02:49:30.150509","2026-04-06T06:53:15.947691",[140,145,150,155,160,164],{"id":141,"question_zh":142,"answer_zh":143,"source_url":144},3316,"API 密钥验证失败或返回 False 怎么办？","请检查是否设置了 `NIXTLA_BASE_URL` 环境变量。如果未设置，默认值可能与实际 API 地址不匹配。确保账户处于活跃状态且拥有足够的积分。如果客户端初始化成功但验证失败，请尝试在其他环境中运行以排除本地环境问题。","https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fissues\u002F396",{"id":146,"question_zh":147,"answer_zh":148,"source_url":149},3317,"遇到 'Too many requests' (429) 错误如何解决？","这表示已达到请求限制。新版本应该能修复此问题。建议前往 dashboard.nixtla.io 创建试用账户并开始使用 TimeGPT。如果问题持续，可发送邮件至 ops@nixtla.io 寻求支持。","https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fissues\u002F145",{"id":151,"question_zh":152,"answer_zh":153,"source_url":154},3318,"如何使 nixtlats 兼容 Pydantic V2？","该兼容性问题已在版本 0.5.0 中修复。请将 nixtlats 升级到 0.5.0 或更高版本，即可解决与 Pydantic v2 的冲突。","https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fissues\u002F228",{"id":156,"question_zh":157,"answer_zh":158,"source_url":159},3319,"关于时间序列 Transformer 模型性能的评价是否存在矛盾？","这取决于具体的上下文和定义的成功标准（如准确性、模型复杂度等）。通用的回答是‘视情况而定’。虽然通用 Transformer 在某些场景下表现不一，但 Nixtla 开发的模型针对时间序列进行了优化，因此在特定实现下效果良好。","https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fissues\u002F422",{"id":161,"question_zh":162,"answer_zh":163,"source_url":144},3320,"是否需要手动设置 NIXTLA_BASE_URL 环境变量？","如果不设置该变量，系统会默认使用 curl 命令中提供的 URL。如果在 Python 客户端中运行正常但在终端中异常，或者反之，可能需要显式设置 `NIXTLA_BASE_URL` 环境变量以确保地址一致。",{"id":165,"question_zh":166,"answer_zh":167,"source_url":149},3321,"如何获取 TimeGPT 的试用账户或访问权限？","所有用户都可以在 dashboard.nixtla.io 上创建试用账户并立即开始使用 TimeGPT。如果遇到访问限制或需要更多支持，可以联系 ops@nixtla.io。",[169,174,179,184,189,194,199,204,209,214,219,224,229,234,239,244,249,254,259,264],{"id":170,"version":171,"summary_zh":172,"released_at":173},102871,"v0.7.4.dev1","- [FIX] Get request prevents raising JSONDecodeError() directly when fail to parse Json() @JQGoh (#790)\r\n\r\n## Dependencies\r\n\r\n- chore(deps): bump release-drafter\u002Frelease-drafter from 6.4.0 to 7.1.1 in the ci-dependencies group @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#785)\r\n- chore(deps): bump cryptography from 46.0.5 to 46.0.6 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#789)\r\n- chore(deps): bump pygments from 2.19.2 to 2.20.0 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#788)\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.7.4.dev0...v0.7.4.dev1","2026-04-01T02:32:37",{"id":175,"version":176,"summary_zh":177,"released_at":178},102880,"v0.6.7.dev2","## What's Changed\r\n* Adjust model input size validation logic, add devcontainer by @goodwanghan in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F632\r\n* Update package version by @goodwanghan in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F633\r\n\r\n## New Contributors\r\n* @goodwanghan made their first contribution in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F632\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.6.7.dev1...v0.6.7.dev2","2025-04-14T04:23:01",{"id":180,"version":181,"summary_zh":182,"released_at":183},102872,"v0.7.4.dev0","## New Features\r\n- [FEAT] Support for categorical features @marcopeix (#760)\r\n- [FEAT] Support missing values @marcopeix (#757)\r\n\r\n## Documentations\r\n- [chore] Update documentation to use `S_df` param @gee-senbong (#786)\r\n \r\n## Improvements\r\n- [chore] Pre-release 0.7.4.dev0 @JQGoh (#787)\r\n- [FEAT] Add model info to request header @JQGoh (#784)\r\n- feat: add PostHog apiHost for ad-blocker bypass @loama (#782)\r\n- [chore] cleanup @deven367 (#778)\r\n- [chore] Clean up Snowflake integration tests and remove stale CI secrets @gee-senbong (#780)\r\n- [fix] Revise snowflake integration @gee-senbong (#777)\r\n- [chore] fix permissions for the workflows @deven367 (#775)\r\n- [chore] Revert `environment: secrets` changes & cleanup finetuned models @gee-senbong (#774)\r\n- fix: fix a bug & broken tests due to recent package upgrades @gee-senbong (#766)\r\n- Update copy for dashboard.nixtla.io to dashboard @rslmrn (#765)\r\n- Page context menu cta now at the level of the eyebrow @rslmrn (#763)\r\n- Replace dashboard.nixtla.io links with free-trial UTM URLs @loama (#762)\r\n- feat: upgrade styles to match with the ones of the new design @rslmrn (#761)\r\n\r\n## Dependencies\r\n- chore(deps): bump requests from 2.32.5 to 2.33.0 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#783)\r\n- chore(deps): bump release-drafter\u002Frelease-drafter from 6.2.0 to 6.4.0 in the ci-dependencies group @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#764)\r\n- chore(deps): bump pyjwt from 2.11.0 to 2.12.0 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#776)\r\n- chore(deps): bump orjson from 3.11.3 to 3.11.6 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#773)\r\n- chore(deps): bump black from 25.9.0 to 26.3.1 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#772)\r\n- chore(deps): bump tornado from 6.5.2 to 6.5.5 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#771)\r\n- chore(deps): bump werkzeug from 3.1.5 to 3.1.6 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#768)\r\n- chore(deps): bump flask from 3.1.2 to 3.1.3 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#769)\r\n- chore(deps): bump ray from 2.52.1 to 2.54.0 @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#758)\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.7.3...v0.7.4.dev0","2026-03-30T16:55:18",{"id":185,"version":186,"summary_zh":187,"released_at":188},102873,"v0.7.3","## New features\r\n- Snowflake integration @gee-senbong (#722)\r\n\r\n## Improvements\r\n- [CHORE] Solves dependabot alerts @nasaul (#746)\r\n- [CHORE] Bumps packages versions @nasaul (#741)\r\n- chore(deps): Deprecate Python 3.9 & bump `fugue` to 0.9.6 @gee-senbong (#737)\r\n- [Chore] Pin pandas \u003C 3.0.0 @JQGoh (#729)\r\n\r\n## Documentation\r\n- [docs] update icons across documentation for consistency and clarity @deven367 (#727)\r\n- [seo] specify icon library @deven367 (#726)\r\n- Remove Amplitude integration @loama (#725)\r\n- DOCS: update support email @loama (#691)\r\n- [docs] add redirects to fix broken links from `intercom` @deven367 (#723)\r\n- Remove marimo iframe from why_timegpt.mdx @loama (#724)\r\n- GenAI banner @loama (#721)\r\n- clean up @deven367 (#719)\r\n- docs: instructions for TimeGPT-2 and cleanup @ngupta23 (#717)\r\n- new banner for timegpt-2.1 @deven367 (#718)\r\n\r\n## Dependencies\r\n- chore(deps): bump werkzeug from 3.1.3 to 3.1.5 @dependabot (#735)\r\n- chore(deps): bump fonttools from 4.60.1 to 4.60.2 @dependabot (#734)\r\n- chore(deps): bump aiohttp from 3.13.1 to 3.13.3 @dependabot (#733)\r\n- chore(deps): bump urllib3 from 2.5.0 to 2.6.3 @dependabot (#732)\r\n- chore(deps-dev): bump setuptools from 69.5.1 to 78.1.1 @dependabot (#731)\r\n- chore(deps): bump ray from 2.20.0 to 2.52.1 @dependabot (#730)","2026-02-13T10:25:37",{"id":190,"version":191,"summary_zh":192,"released_at":193},102874,"v0.7.2","- Prepare for 0.7.2 release @goodwanghan (#715)\r\n- [Chore] Omit model name check @JQGoh (#714)\r\n- [FIX] Deprecate insample endpoint @elephaint (#712)\r\n- Update docs.json @loama (#710)\r\n- [DOCS] Add MLFLow tutorial @marcopeix (#713)\r\n","2025-12-10T22:36:56",{"id":195,"version":196,"summary_zh":197,"released_at":198},102875,"v0.7.1","## New features\r\n- [FEAT] model_parameters argument @JQGoh (#681)\r\n- [FEAT]: model_parameters support in cross_validation method  @JQGoh (#694)\r\n- [FEAT] Reduce finetune requirements @elephaint (#680)\r\n- [FEAT] Multivariate endpoint @elephaint (#708)\r\n\r\n## Bug fixes\r\n- Fix codespace and tests @goodwanghan (#693)\r\n\r\n## Documentation\r\n- Banner for timegpt-2.0 @deven367 (#689)\r\n- Fix table render @deven367 (#706)\r\n- Fix use cases links @loama (#670)\r\n- Third party licenses @deven367 (#704)\r\n- Skip test if only docs are updated @deven367 (#702)\r\n- Add intercom integration @deven367 (#701)\r\n- Remove `deploy-readme` workflow @deven367 (#685)\r\n- Edit use cases @khuyentran1401 (#698)\r\n- Edit real time anomaly detection @khuyentran1401 (#695)\r\n- Rewrite forecasting at scale @khuyentran1401 (#692)\r\n- Rewrite longhorizon @khuyentran1401 (#690)\r\n- Docs improve seo of forecasting quickstart @khuyentran1401 (#688)\r\n- Fix typo in the redirects @deven367 (#687)\r\n- Have redirects point to the correct url @deven367 (#686)\r\n- Skip tests when only docs are updated @deven367 (#679)\r\n- Improve cross validation seo @khuyentran1401 (#683)\r\n- Optimize bitcoin prediction @khuyentran1401 (#682)\r\n- Added API reference to docs @mergenthaler (#672)\r\n- Update image paths in documentation to fix broken images @khuyentran1401 (#678)\r\n- Styling @loama (#671)\r\n- Fixed broken Links, Exclude pending content and add contextual options to documentation @mergenthaler (#675)\r\n- Failing docs ci @deven367 (#667)\r\n- Failing docs ci @deven367 (#666)","2025-11-19T11:49:17",{"id":200,"version":201,"summary_zh":202,"released_at":203},102876,"v0.7.1.dev0","- Release of 0.7.1.dev0 @JQGoh (#699)\r\n\r\n## New Features\r\n- [FEAT]: model_parameters support in cross_validation method  @JQGoh (#694)\r\n- [FEAT] Reduce finetune requirements @elephaint (#680)\r\n- [FEAT] model_parameters argument @JQGoh (#681)\r\n\r\n## Documentation\r\n- Docs edit use cases @khuyentran1401 (#698)\r\n- Docs edit real time anomaly detection @khuyentran1401 (#695)\r\n- Docs rewrite forecasting at scale @khuyentran1401 (#692)\r\n- Docs rewrite longhorizon @khuyentran1401 (#690)\r\n- banner for timegpt-2.0 @deven367 (#689)\r\n- Docs improve seo of forecasting quickstart @khuyentran1401 (#688)\r\n- [docs] fix typo in the redirects @deven367 (#687)\r\n- [docs] have redirects point to the correct url @deven367 (#686)\r\n- [docs] skip tests when only docs are updated @deven367 (#679)\r\n- Docs improve cross validation seo @khuyentran1401 (#683)\r\n- Docs optimize bitcoin prediction @khuyentran1401 (#682)\r\n- feat: added API reference to docs @mergenthaler (#672)\r\n- Update image paths in documentation to fix broken images @khuyentran1401 (#678)\r\n- [DOCS] styling @loama (#671)\r\n- docs: Fixed broken Links, Exclude pending content and add contextual options to documentation @mergenthaler (#675)\r\n- fix use cases links @loama (#670)\r\n- Failing docs ci @deven367 (#667)\r\n- Failing docs ci @deven367 (#666)\r\n\r\n## Dependency and tests\r\n- Fix codespace and tests @goodwanghan (#693)\r\n\r\n## New Contributors\r\n* @khuyentran1401 made their first contribution in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F678\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.7.0...v0.7.1.dev0","2025-10-31T09:08:44",{"id":205,"version":206,"summary_zh":207,"released_at":208},102877,"v0.7.0","## What's Changed\r\n* Release 0.7.0 by @goodwanghan in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F665\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.6.8.dev1...v0.7.0","2025-10-31T03:27:04",{"id":210,"version":211,"summary_zh":212,"released_at":213},102878,"v0.6.8.dev1","## New Features\r\n* [FEAT] Reduce minimum required size for finetuning and fix testing failures by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F640\r\n\r\n\r\n## Bug Fixes\r\n* FIX: Add closing steps tag by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F659\r\n\r\n## Documentation\r\n* feat: new date features tutorial by @ngupta23 in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F653\r\n* Migrate docs by @loama in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F656\r\n* DOCS: Faq and about fixes by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F657\r\n* add posthog to analytics in mintlify by @loama in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F660\r\n* New documentation pipeline by @deven367 in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F663\r\n* test for staging docs by @deven367 in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F664\r\n\r\n## Dependencies\r\n* Remove nbdev for core developments and use pytest for tests by @JQGoh in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F647\r\n* CHORE: add nodejs to codespace by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F658\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.6.7.dev3...v0.6.8.dev1","2025-10-31T02:59:40",{"id":215,"version":216,"summary_zh":217,"released_at":218},102879,"v0.6.7.dev3","## What's Changed\r\n* Relax supported model check by @goodwanghan in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F634\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.6.7.dev2...v0.6.7.dev3","2025-04-29T22:54:07",{"id":220,"version":221,"summary_zh":222,"released_at":223},102881,"v0.6.7.dev1","- chore: v0.6.7.dev1 @Yibei990826 (#630)\r\n- refactor: Replace httpx.Client with _make_client @Yibei990826 (#629)\r\n- [DOC] add faq on simple data forecast @Yibei990826 (#588)\r\n- [doc] Add Meeting Link to Pricing @Yibei990826 (#626)\r\n- feat: initial version of the audit & clean data feature @ngupta23 (#611)\r\n- fix: broken pricing link @ngupta23 (#625)\r\n- [FEAT] Temporal Hierarchical with TimeGPT @elephaint (#623)\r\n- doc: unified wording on pricing page @ngupta23 (#624)\r\n- docs: updated pricing to remove Azure and Free plan. @ngupta23 (#617)\r\n- doc: updated pricing @ngupta23 (#615)\r\n- fix: for test failure on windows 3.9 @ngupta23 (#616)\r\n- doc: added info about team seats @ngupta23 (#603)\r\n- docs: add tutorial on custom frequencies @MMenchero (#597)\r\n\r\n## New Features\r\n\r\n- feat: get single finetuned model @jmoralez (#610)\r\n\r\n## Bug Fixes\r\n\r\n- misc: Fix typing for client timeout @adamantike (#621)\r\n\r\n## Documentation\r\n\r\n- [FIX] Docs incorrect mailto links @elephaint (#574)\r\n- docs: update changelog @jmoralez (#595)\r\n- [FIX] Change html table to md format @Yibei990826 (#596)\r\n\r\n## Dependencies\r\n\r\n- fix: lint @ngupta23 (#620)\r\n- chore(deps): bump actions\u002Fsetup-python from 5.3.0 to 5.4.0 in the ci-dependencies group @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#600)\r\n- chore(deps): bump pypa\u002Fgh-action-pypi-publish from 1.12.3 to 1.12.4 in the ci-dependencies group @[dependabot[bot]](https:\u002F\u002Fgithub.com\u002Fapps\u002Fdependabot) (#598)\r\n","2025-04-11T06:28:38",{"id":225,"version":226,"summary_zh":227,"released_at":228},102882,"v0.6.6","## New Features\r\n\r\n- FEAT: online anomaly detection @marcopeix (#546)\r\n\r\n## Bug Fixes\r\n\r\n- [FIX] exogenous features with partitioning @elephaint (#591)\r\n","2025-01-20T20:59:04",{"id":230,"version":231,"summary_zh":232,"released_at":233},102883,"v0.6.5","## New Features\r\n- feat: saving and re-using fine-tuned models @jmoralez (#562)\r\n- feat: add refit argument to cross_validation @jmoralez (#556)\r\n- feat: zstd-compress requests over 1MB @jmoralez (#559)\r\n\r\n## Documentation\r\n- docs: updated to include `fill_gaps` capability for multiple time series @ngupta23 (#580)\r\n- docs: cleanup @ngupta23 (#578)\r\n- feat: add vn1 competition experiment @AzulGarza (#563)\r\n- docs: FAQ | formatting fix @ngupta23 (#569)\r\n- [DOCS] Fixes two docs issues @elephaint (#570)\r\n- docs: updated FAQs @ngupta23 (#564)\r\n- [FIX] Img links @elephaint (#558)\r\n- [FIX] Img link @elephaint (#557)\r\n- DOC: separate tutorial for finetune_depth @marcopeix (#547)\r\n- fix broken links and image @Yibei990826 (#551)\r\n- [FIX] Electricity short @elephaint (#555)\r\n- [FIX] Electricity dataset on tutorials @elephaint (#554)\r\n\r\n## Enhancement\r\n- enh: use GET request for model_params @jmoralez (#576)\r\n- enh: update validate_api_key to use GET request @jmoralez (#573)\r\n","2025-01-02T20:26:01",{"id":235,"version":236,"summary_zh":237,"released_at":238},102884,"v0.6.4","## New Features\r\n\r\n- feat: support custom and integer frequencies @jmoralez (#532)\r\n- feat: usage method @jmoralez (#548)\r\n- feat: add hist_exog_list argument to cross_validation @jmoralez (#534)\r\n\r\n## Documentation\r\n\r\n- Add Why TimeGPT notebook @Yibei990826 (#537)\r\n- [FIX] Pricing links incorrect @elephaint (#544)\r\n- [FIX] Update contributing guide docs @elephaint (#539)\r\n- fix: cleaned the FAQ related to API validation @ngupta23 (#538)\r\n- docs: instructions on how to contribute @ngupta23 (#526)\r\n- docs: broken links in intro notebook @ngupta23 (#527)\r\n- docs: updated dashboard image to latest dashboard @ngupta23 (#528)\r\n","2024-12-02T22:25:41",{"id":240,"version":241,"summary_zh":242,"released_at":243},102885,"v0.6.3","## Bug Fixes\r\n\r\n- fix(deps): bump utils @jmoralez (#524)\r\n","2024-11-04T19:01:05",{"id":245,"version":246,"summary_zh":247,"released_at":248},102886,"v0.6.2","## New Features\r\n\r\n- FEAT: Add finetune_depth parameter @marcopeix (#471)\r\n\r\n## Breaking Change\r\n\r\n- breaking: raise error for gaps in series @jmoralez (#504)\r\n-  breaking: add `hist_exog_list` argument to forecast @jmoralez (#505)\r\n\r\n## Bug Fixes\r\n\r\n- fix: convert level to quantiles in historic forecast @jmoralez (#510)\r\n- fix: reduce test time @elephaint (#487)\r\n- fix(ci): increase tolerance in tests @jmoralez (#503)\r\n- fix: update docs to add hist exog list instead of previous behaviour @AzulGarza (#520)\r\n\r\n## Documentation\r\n\r\n- Updating Azure getting started guide.  @tracykteal (#508)\r\n- updating Azure and trial information @tracykteal (#500)\r\n- Adding a pricing page to our docs @tracykteal (#502)\r\n- docs: add finetune_depth to tutorial on improving accuracy @marcopeix (#497)\r\n- docs: add improve accuracy content to special topics page @Yibei990826 (#496)\r\n- docs: add notebook tutorial for improve forecast accuracy @Yibei990826 (#495)\r\n- docs: update energy demand forecast tutorial @Yibei990826 (#499)\r\n\r\n## Enhancement\r\n\r\n- enh: warn when overriding model in Azure endpoint @jmoralez (#511)\r\n- enh: improve short series error message @jmoralez (#494)\r\n","2024-11-04T04:35:31",{"id":250,"version":251,"summary_zh":252,"released_at":253},102887,"v0.6.1","# New Features\r\n* feat: support newer azure deployments by @jmoralez in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F478\r\n\r\n# Documentation\r\n* [DOCS] Add API call count FAQ by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F449\r\n* Adding Azure getting started doc by @tracykteal in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F474\r\n\r\n# Other\r\n* chore: remove autogenerated code by @jmoralez in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F468\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.6.0...v0.6.1","2024-10-01T03:04:26",{"id":255,"version":256,"summary_zh":257,"released_at":258},102888,"v0.6.0","# Breaking changes\r\n\r\n* Removed `TimeGPT` class in favor of `NixtlaClient`.\r\n* Removed `NixtlaClient.validate_token` method in favor of `NixtlaClient.validate_api_key`.\r\n* Removed `short-horizon` and `long-horizon` models in favor of `timegpt-1` and `timegpt-1-long-horizon` respectively.\r\n* Removed `fewshot_steps` and `fewshot_loss` in favor of `finetune_steps` and `finetune_loss` respectively.\r\n* Removed `TIMEGPT_TOKEN` environment variable in favor of `NIXTLA_API_KEY`.\r\n* Timestamps in the `time_col` preserve their type in the outputs (timestamp), previously they were cast to string.\r\n* Gaps in series are not filled anymore, in line with [our documentation](https:\u002F\u002Fnixtlaverse.nixtla.io\u002Fnixtla\u002Fdocs\u002Fgetting-started\u002Fdata_requirements.html#important-considerations). Please use the [fill_gaps function](https:\u002F\u002Fnixtlaverse.nixtla.io\u002Futilsforecast\u002Fpreprocessing.html#fill-gaps) if you require to do so.\r\n* `NixtlaClient.weights_x` is now a list of lists if `num_partitions != None`, where each element corresponds to the weights for a specific partition.\r\n\r\n# Features\r\n* feat: call v2 endpoints by @jmoralez in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F429\r\n* feat: orjson serialization by @jmoralez in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F452\r\n* [FEAT] Add historical exogenous by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F453\r\n* FEAT: Add feature contributions as attribute and tutorial on how to use by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F448\r\n* [FEAT] - Raise error when payload is too large and suggest number of partitions by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F456\r\n\r\n# Bug fixes\r\n* fix: removed data input restriction during cross validation finetune by @Yibei990826 in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F426\r\n\r\n# Enhancements\r\n* feat: use TypeVar for dataframes by @jmoralez in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F443\r\n* feat: set supported_models using base_url by @jmoralez in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F450\r\n* feat: suggest earlier version for azure endpoints by @jmoralez in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F457\r\n\r\n# Documentation\r\n* updating excel docs with some more information by @tracykteal in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F414\r\n* updating FAQ for fee information and Azure being available by @tracykteal in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F412\r\n* Update 03_excel_addin.ipynb by @tracykteal in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F416\r\n* [DOC] - Rephrase the sentence for specific loss function by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F436\r\n* docs(readme): getting started by @mergenthaler in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F438\r\n* Adding activate trial link in intro by @tracykteal in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F417\r\n* [DOC] - Add instructions on saving figures when not working in notebooks by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F454\r\n* [BUG]: close img tag in introduction notebook by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F441\r\n* docs: add polars quickstart by @jmoralez in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F447\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.5.2...v0.6.0","2024-09-03T20:33:48",{"id":260,"version":261,"summary_zh":262,"released_at":263},102889,"v0.5.2","## What's Changed\r\n\r\n### Features\r\n\r\n* Use case on missing values by @MMenchero in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F326\r\n* Electricity demand use case by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F321\r\n* [FEAT] What if - pricing in retail scenario by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F340\r\n* Update the middleware endpoint to use the new one by @loama in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F407\r\n* Update excel add in docs to reflect changes in the plugin by @loama in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F406\r\n* feat: improve readability by @mergenthaler in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F367\r\n* [DOC] Add azure callouts to use cases by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F357\r\n* [DOC] - Add links and callouts by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F355\r\n* [DOC] Model callouts and descriptions by @MMenchero in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F359\r\n* feat: add foundation ts arena by @AzulGarza in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F378\r\n* feat: Added nixtlar to TimeGPT docs  by @MMenchero in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F393\r\n\r\n### Fixes\r\n\r\n* [FIX] SDK Reference by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F342\r\n* [FIX] Raise warning on missing X_df when df has exogenous by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F295\r\n* fix: add jupyter lab to dev deps by @mergenthaler in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F350\r\n* [FIX] Hotfix for old SDK link removal by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F352\r\n* [Hotfix] Minimum data requirements by @MMenchero in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F364\r\n* fix: note on morai improvements by @mergenthaler in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F395\r\n* fix: Clarify how tokens are counted for AzureAI by @MMenchero in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F401\r\n* Hotfix for links in documentation by @MMenchero in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F341\r\n* [FIX] Colab-flag by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F349\r\n* Minor correction to tutorial by @MMenchero in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F394\r\n\r\n### Chore\r\n* Readme update by @mergenthaler in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F346\r\n* Add Fern badge to README.md by @dannysheridan in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F185\r\n* [DOC] - Reduce code blocks in capabilties by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F339\r\n* Better intro\u002Fwelcome page for doc by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F343\r\n* a few updates to getting started by @tracykteal in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F344\r\n* fix: note on chronos pr by @mergenthaler in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F390\r\n* fix: Change how data is downloaded for Bitcoin tutorial by @MMenchero in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F391\r\n* Removed a wrong sentence in README.md by @B-Deforce in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F372\r\n\r\n## New Contributors\r\n* @dannysheridan made their first contribution in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F185\r\n* @B-Deforce made their first contribution in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F372\r\n* @loama made their first contribution in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F407\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.5.1...v0.5.2","2024-07-05T11:30:04",{"id":265,"version":266,"summary_zh":267,"released_at":268},102890,"v0.5.1","## What's Changed\r\n* [DOCS] - Add a welcome\u002Fintroduction page to the documentation by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F319\r\n* [FIX] Doc fixes by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F324\r\n* Add capabilities notebooks by @marcopeix in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F312\r\n* [DOCS] Fix structure by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F329\r\n* [FIX] Mintlify prefixes, readme file dirs by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F333\r\n* [FIX] Doc fixes by @elephaint in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F334\r\n* fix: add pyreadr as dev dep by @AzulGarza in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F336\r\n* fix: update distributed to computing at scale by @AzulGarza in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F337\r\n* fix: add model to cross validation path call by @AzulGarza in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F335\r\n* v0.5.1 by @AzulGarza in https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fpull\u002F338\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FNixtla\u002Fnixtla\u002Fcompare\u002Fv0.5.0...v0.5.1","2024-05-06T19:07:25"]