[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-ollama--ollama-python":3,"tool-ollama--ollama-python":65},[4,17,27,35,48,57],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"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",150037,2,"2026-04-10T23:33:47",[13,14,15],"开发框架","Agent","语言模型","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},4487,"LLMs-from-scratch","rasbt\u002FLLMs-from-scratch","LLMs-from-scratch 是一个基于 PyTorch 的开源教育项目，旨在引导用户从零开始一步步构建一个类似 ChatGPT 的大型语言模型（LLM）。它不仅是同名技术著作的官方代码库，更提供了一套完整的实践方案，涵盖模型开发、预训练及微调的全过程。\n\n该项目主要解决了大模型领域“黑盒化”的学习痛点。许多开发者虽能调用现成模型，却难以深入理解其内部架构与训练机制。通过亲手编写每一行核心代码，用户能够透彻掌握 Transformer 架构、注意力机制等关键原理，从而真正理解大模型是如何“思考”的。此外，项目还包含了加载大型预训练权重进行微调的代码，帮助用户将理论知识延伸至实际应用。\n\nLLMs-from-scratch 特别适合希望深入底层原理的 AI 开发者、研究人员以及计算机专业的学生。对于不满足于仅使用 API，而是渴望探究模型构建细节的技术人员而言，这是极佳的学习资源。其独特的技术亮点在于“循序渐进”的教学设计：将复杂的系统工程拆解为清晰的步骤，配合详细的图表与示例，让构建一个虽小但功能完备的大模型变得触手可及。无论你是想夯实理论基础，还是为未来研发更大规模的模型做准备",90106,3,"2026-04-06T11:19:32",[15,26,14,13],"图像",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":10,"last_commit_at":33,"category_tags":34,"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",[13,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":10,"last_commit_at":41,"category_tags":42,"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。",85092,"2026-04-10T11:13:16",[26,43,44,45,14,46,15,13,47],"数据工具","视频","插件","其他","音频",{"id":49,"name":50,"github_repo":51,"description_zh":52,"stars":53,"difficulty_score":54,"last_commit_at":55,"category_tags":56,"status":16},5784,"funNLP","fighting41love\u002FfunNLP","funNLP 是一个专为中文自然语言处理（NLP）打造的超级资源库，被誉为\"NLP 民工的乐园”。它并非单一的软件工具，而是一个汇集了海量开源项目、数据集、预训练模型和实用代码的综合性平台。\n\n面对中文 NLP 领域资源分散、入门门槛高以及特定场景数据匮乏的痛点，funNLP 提供了“一站式”解决方案。这里不仅涵盖了分词、命名实体识别、情感分析、文本摘要等基础任务的标准工具，还独特地收录了丰富的垂直领域资源，如法律、医疗、金融行业的专用词库与数据集，甚至包含古诗词生成、歌词创作等趣味应用。其核心亮点在于极高的全面性与实用性，从基础的字典词典到前沿的 BERT、GPT-2 模型代码，再到高质量的标注数据和竞赛方案，应有尽有。\n\n无论是刚刚踏入 NLP 领域的学生、需要快速验证想法的算法工程师，还是从事人工智能研究的学者，都能在这里找到急需的“武器弹药”。对于开发者而言，它能大幅减少寻找数据和复现模型的时间；对于研究者，它提供了丰富的基准测试资源和前沿技术参考。funNLP 以开放共享的精神，极大地降低了中文自然语言处理的开发与研究成本，是中文 AI 社区不可或缺的宝藏仓库。",79857,1,"2026-04-08T20:11:31",[15,43,46],{"id":58,"name":59,"github_repo":60,"description_zh":61,"stars":62,"difficulty_score":54,"last_commit_at":63,"category_tags":64,"status":16},6590,"gpt4all","nomic-ai\u002Fgpt4all","GPT4All 是一款让普通电脑也能轻松运行大型语言模型（LLM）的开源工具。它的核心目标是打破算力壁垒，让用户无需依赖昂贵的显卡（GPU）或云端 API，即可在普通的笔记本电脑和台式机上私密、离线地部署和使用大模型。\n\n对于担心数据隐私、希望完全掌控本地数据的企业用户、研究人员以及技术爱好者来说，GPT4All 提供了理想的解决方案。它解决了传统大模型必须联网调用或需要高端硬件才能运行的痛点，让日常设备也能成为强大的 AI 助手。无论是希望构建本地知识库的开发者，还是单纯想体验私有化 AI 聊天的普通用户，都能从中受益。\n\n技术上，GPT4All 基于高效的 `llama.cpp` 后端，支持多种主流模型架构（包括最新的 DeepSeek R1 蒸馏模型），并采用 GGUF 格式优化推理速度。它不仅提供界面友好的桌面客户端，支持 Windows、macOS 和 Linux 等多平台一键安装，还为开发者提供了便捷的 Python 库，可轻松集成到 LangChain 等生态中。通过简单的下载和配置，用户即可立即开始探索本地大模型的无限可能。",77307,"2026-04-11T06:52:37",[15,13],{"id":66,"github_repo":67,"name":68,"description_en":69,"description_zh":70,"ai_summary_zh":70,"readme_en":71,"readme_zh":72,"quickstart_zh":73,"use_case_zh":74,"hero_image_url":75,"owner_login":76,"owner_name":77,"owner_avatar_url":78,"owner_bio":79,"owner_company":80,"owner_location":80,"owner_email":81,"owner_twitter":76,"owner_website":82,"owner_url":83,"languages":84,"stars":89,"forks":90,"last_commit_at":91,"license":92,"difficulty_score":10,"env_os":93,"env_gpu":94,"env_ram":95,"env_deps":96,"category_tags":101,"github_topics":102,"view_count":10,"oss_zip_url":80,"oss_zip_packed_at":80,"status":16,"created_at":104,"updated_at":105,"faqs":106,"releases":135},4960,"ollama\u002Follama-python","ollama-python","Ollama Python library","ollama-python 是专为 Python 开发者打造的官方库，旨在让集成本地或云端大语言模型变得简单高效。它解决了在 Python 项目中调用 AI 模型时配置复杂、代码冗余的痛点，只需几行代码即可实现与 Ollama 服务的无缝对接。\n\n无论是需要在本地运行轻量级模型，还是希望调用云端超大参数模型（如千亿级参数），ollama-python 都能轻松胜任。它支持同步和异步两种编程模式，开发者可以根据项目需求灵活选择；同时提供流式响应功能，让 AI 生成的内容能够实时逐字输出，极大提升了交互体验。此外，它还允许自定义 HTTP 客户端配置，方便高级用户进行深度定制。\n\n这款工具特别适合 Python 开发者、AI 研究人员以及需要快速原型验证的技术团队使用。如果你正在构建聊天机器人、智能助手或任何需要自然语言处理能力的应用，ollama-python 能让你专注于业务逻辑，而无需担忧底层模型调用的复杂性。配合 Ollama 本地服务或云端 API，你可以自由切换模型资源，在保持开发流程一致性的同时，灵活应对不同规模的计算需求。","# Ollama Python Library\n\nThe Ollama Python library provides the easiest way to integrate Python 3.8+ projects with [Ollama](https:\u002F\u002Fgithub.com\u002Follama\u002Follama).\n\n## Prerequisites\n\n- [Ollama](https:\u002F\u002Follama.com\u002Fdownload) should be installed and running\n- Pull a model to use with the library: `ollama pull \u003Cmodel>` e.g. `ollama pull gemma3`\n  - See [Ollama.com](https:\u002F\u002Follama.com\u002Fsearch) for more information on the models available.\n\n## Install\n\n```sh\npip install ollama\n```\n\n## Usage\n\n```python\nfrom ollama import chat\nfrom ollama import ChatResponse\n\nresponse: ChatResponse = chat(model='gemma3', messages=[\n  {\n    'role': 'user',\n    'content': 'Why is the sky blue?',\n  },\n])\nprint(response['message']['content'])\n# or access fields directly from the response object\nprint(response.message.content)\n```\n\nSee [_types.py](ollama\u002F_types.py) for more information on the response types.\n\n## Streaming responses\n\nResponse streaming can be enabled by setting `stream=True`.\n\n```python\nfrom ollama import chat\n\nstream = chat(\n    model='gemma3',\n    messages=[{'role': 'user', 'content': 'Why is the sky blue?'}],\n    stream=True,\n)\n\nfor chunk in stream:\n  print(chunk['message']['content'], end='', flush=True)\n```\n\n## Cloud Models\n\nRun larger models by offloading to Ollama’s cloud while keeping your local workflow.\n\n- Supported models: `deepseek-v3.1:671b-cloud`, `gpt-oss:20b-cloud`, `gpt-oss:120b-cloud`, `kimi-k2:1t-cloud`, `qwen3-coder:480b-cloud`, `kimi-k2-thinking` See [Ollama Models - Cloud](https:\u002F\u002Follama.com\u002Fsearch?c=cloud) for more information\n\n### Run via local Ollama\n\n1) Sign in (one-time):\n\n```\nollama signin\n```\n\n2) Pull a cloud model:\n\n```\nollama pull gpt-oss:120b-cloud\n```\n\n3) Make a request:\n\n```python\nfrom ollama import Client\n\nclient = Client()\n\nmessages = [\n  {\n    'role': 'user',\n    'content': 'Why is the sky blue?',\n  },\n]\n\nfor part in client.chat('gpt-oss:120b-cloud', messages=messages, stream=True):\n  print(part.message.content, end='', flush=True)\n```\n\n### Cloud API (ollama.com)\n\nAccess cloud models directly by pointing the client at `https:\u002F\u002Follama.com`.\n\n1) Create an API key from [ollama.com](https:\u002F\u002Follama.com\u002Fsettings\u002Fkeys) , then set:\n\n```\nexport OLLAMA_API_KEY=your_api_key\n```\n\n2) (Optional) List models available via the API:\n\n```\ncurl https:\u002F\u002Follama.com\u002Fapi\u002Ftags\n```\n\n3) Generate a response via the cloud API:\n\n```python\nimport os\nfrom ollama import Client\n\nclient = Client(\n    host='https:\u002F\u002Follama.com',\n    headers={'Authorization': 'Bearer ' + os.environ.get('OLLAMA_API_KEY')}\n)\n\nmessages = [\n  {\n    'role': 'user',\n    'content': 'Why is the sky blue?',\n  },\n]\n\nfor part in client.chat('gpt-oss:120b', messages=messages, stream=True):\n  print(part.message.content, end='', flush=True)\n```\n\n## Custom client\nA custom client can be created by instantiating `Client` or `AsyncClient` from `ollama`.\n\nAll extra keyword arguments are passed into the [`httpx.Client`](https:\u002F\u002Fwww.python-httpx.org\u002Fapi\u002F#client).\n\n```python\nfrom ollama import Client\nclient = Client(\n  host='http:\u002F\u002Flocalhost:11434',\n  headers={'x-some-header': 'some-value'}\n)\nresponse = client.chat(model='gemma3', messages=[\n  {\n    'role': 'user',\n    'content': 'Why is the sky blue?',\n  },\n])\n```\n\n## Async client\n\nThe `AsyncClient` class is used to make asynchronous requests. It can be configured with the same fields as the `Client` class.\n\n```python\nimport asyncio\nfrom ollama import AsyncClient\n\nasync def chat():\n  message = {'role': 'user', 'content': 'Why is the sky blue?'}\n  response = await AsyncClient().chat(model='gemma3', messages=[message])\n\nasyncio.run(chat())\n```\n\nSetting `stream=True` modifies functions to return a Python asynchronous generator:\n\n```python\nimport asyncio\nfrom ollama import AsyncClient\n\nasync def chat():\n  message = {'role': 'user', 'content': 'Why is the sky blue?'}\n  async for part in await AsyncClient().chat(model='gemma3', messages=[message], stream=True):\n    print(part['message']['content'], end='', flush=True)\n\nasyncio.run(chat())\n```\n\n## API\n\nThe Ollama Python library's API is designed around the [Ollama REST API](https:\u002F\u002Fgithub.com\u002Follama\u002Follama\u002Fblob\u002Fmain\u002Fdocs\u002Fapi.md)\n\n### Chat\n\n```python\nollama.chat(model='gemma3', messages=[{'role': 'user', 'content': 'Why is the sky blue?'}])\n```\n\n### Generate\n\n```python\nollama.generate(model='gemma3', prompt='Why is the sky blue?')\n```\n\n### List\n\n```python\nollama.list()\n```\n\n### Show\n\n```python\nollama.show('gemma3')\n```\n\n### Create\n\n```python\nollama.create(model='example', from_='gemma3', system=\"You are Mario from Super Mario Bros.\")\n```\n\n### Copy\n\n```python\nollama.copy('gemma3', 'user\u002Fgemma3')\n```\n\n### Delete\n\n```python\nollama.delete('gemma3')\n```\n\n### Pull\n\n```python\nollama.pull('gemma3')\n```\n\n### Push\n\n```python\nollama.push('user\u002Fgemma3')\n```\n\n### Embed\n\n```python\nollama.embed(model='gemma3', input='The sky is blue because of rayleigh scattering')\n```\n\n### Embed (batch)\n\n```python\nollama.embed(model='gemma3', input=['The sky is blue because of rayleigh scattering', 'Grass is green because of chlorophyll'])\n```\n\n### Ps\n\n```python\nollama.ps()\n```\n\n## Errors\n\nErrors are raised if requests return an error status or if an error is detected while streaming.\n\n```python\nmodel = 'does-not-yet-exist'\n\ntry:\n  ollama.chat(model)\nexcept ollama.ResponseError as e:\n  print('Error:', e.error)\n  if e.status_code == 404:\n    ollama.pull(model)\n```\n","# Ollama Python 库\n\nOllama Python 库为将 Python 3.8 及以上版本的项目与 [Ollama](https:\u002F\u002Fgithub.com\u002Follama\u002Follama) 集成提供了最简便的方式。\n\n## 先决条件\n\n- 已安装并运行 [Ollama](https:\u002F\u002Follama.com\u002Fdownload)\n- 拉取一个模型以供库使用：`ollama pull \u003Cmodel>`，例如 `ollama pull gemma3`\n  - 更多可用模型信息，请参阅 [Ollama.com](https:\u002F\u002Follama.com\u002Fsearch)。\n\n## 安装\n\n```sh\npip install ollama\n```\n\n## 使用方法\n\n```python\nfrom ollama import chat\nfrom ollama import ChatResponse\n\nresponse: ChatResponse = chat(model='gemma3', messages=[\n  {\n    'role': 'user',\n    'content': '为什么天空是蓝色的？',\n  },\n])\nprint(response['message']['content'])\n# 或直接从响应对象访问字段\nprint(response.message.content)\n```\n\n有关响应类型的更多信息，请参阅 [_types.py](ollama\u002F_types.py)。\n\n## 流式响应\n\n通过设置 `stream=True` 可启用流式响应。\n\n```python\nfrom ollama import chat\n\nstream = chat(\n    model='gemma3',\n    messages=[{'role': 'user', 'content': '为什么天空是蓝色的？'}],\n    stream=True,\n)\n\nfor chunk in stream:\n  print(chunk['message']['content'], end='', flush=True)\n```\n\n## 云端模型\n\n通过将计算任务卸载到 Ollama 的云端，您可以在保持本地工作流程的同时运行更大的模型。\n\n- 支持的模型：`deepseek-v3.1:671b-cloud`、`gpt-oss:20b-cloud`、`gpt-oss:120b-cloud`、`kimi-k2:1t-cloud`、`qwen3-coder:480b-cloud`、`kimi-k2-thinking`。更多信息请参阅 [Ollama Models - Cloud](https:\u002F\u002Follama.com\u002Fsearch?c=cloud)。\n\n### 通过本地 Ollama 运行\n\n1) 登录（仅需一次）：\n\n```\nollama signin\n```\n\n2) 拉取云端模型：\n\n```\nollama pull gpt-oss:120b-cloud\n```\n\n3) 发起请求：\n\n```python\nfrom ollama import Client\n\nclient = Client()\n\nmessages = [\n  {\n    'role': 'user',\n    'content': '为什么天空是蓝色的？',\n  },\n]\n\nfor part in client.chat('gpt-oss:120b-cloud', messages=messages, stream=True):\n  print(part.message.content, end='', flush=True)\n```\n\n### 云端 API (ollama.com)\n\n通过将客户端指向 `https:\u002F\u002Follama.com`，您可以直接访问云端模型。\n\n1) 从 [ollama.com](https:\u002F\u002Follama.com\u002Fsettings\u002Fkeys) 创建 API 密钥，然后设置：\n\n```\nexport OLLAMA_API_KEY=your_api_key\n```\n\n2) （可选）列出可通过 API 使用的模型：\n\n```\ncurl https:\u002F\u002Follama.com\u002Fapi\u002Ftags\n```\n\n3) 通过云端 API 生成响应：\n\n```python\nimport os\nfrom ollama import Client\n\nclient = Client(\n    host='https:\u002F\u002Follama.com',\n    headers={'Authorization': 'Bearer ' + os.environ.get('OLLAMA_API_KEY')}\n)\n\nmessages = [\n  {\n    'role': 'user',\n    'content': '为什么天空是蓝色的？',\n  },\n]\n\nfor part in client.chat('gpt-oss:120b', messages=messages, stream=True):\n  print(part.message.content, end='', flush=True)\n```\n\n## 自定义客户端\n可以通过实例化 `ollama` 中的 `Client` 或 `AsyncClient` 来创建自定义客户端。\n\n所有额外的关键字参数都会传递给 [`httpx.Client`](https:\u002F\u002Fwww.python-httpx.org\u002Fapi\u002F#client)。\n\n```python\nfrom ollama import Client\nclient = Client(\n  host='http:\u002F\u002Flocalhost:11434',\n  headers={'x-some-header': 'some-value'}\n)\nresponse = client.chat(model='gemma3', messages=[\n  {\n    'role': 'user',\n    'content': '为什么天空是蓝色的？',\n  },\n])\n```\n\n## 异步客户端\n\n`AsyncClient` 类用于发出异步请求。它可以使用与 `Client` 类相同的配置项进行配置。\n\n```python\nimport asyncio\nfrom ollama import AsyncClient\n\nasync def chat():\n  message = {'role': 'user', 'content': '为什么天空是蓝色的？'}\n  response = await AsyncClient().chat(model='gemma3', messages=[message])\n\nasyncio.run(chat())\n```\n\n设置 `stream=True` 会将函数修改为返回 Python 异步生成器：\n\n```python\nimport asyncio\nfrom ollama import AsyncClient\n\nasync def chat():\n  message = {'role': 'user', 'content': '为什么天空是蓝色的？'}\n  async for part in await AsyncClient().chat(model='gemma3', messages=[message], stream=True):\n    print(part['message']['content'], end='', flush=True)\n\nasyncio.run(chat())\n```\n\n## API\n\nOllama Python 库的 API 是围绕 [Ollama REST API](https:\u002F\u002Fgithub.com\u002Follama\u002Follama\u002Fblob\u002Fmain\u002Fdocs\u002Fapi.md) 设计的。\n\n### 聊天\n\n```python\nollama.chat(model='gemma3', messages=[{'role': 'user', 'content': '为什么天空是蓝色的？'}])\n```\n\n### 生成\n\n```python\nollama.generate(model='gemma3', prompt='为什么天空是蓝色的？')\n```\n\n### 列表\n\n```python\nollama.list()\n```\n\n### 显示\n\n```python\nollama.show('gemma3')\n```\n\n### 创建\n\n```python\nollama.create(model='example', from_='gemma3', system=\"你就是超级马里奥中的马里奥。\")\n```\n\n### 复制\n\n```python\nollama.copy('gemma3', 'user\u002Fgemma3')\n```\n\n### 删除\n\n```python\nollama.delete('gemma3')\n```\n\n### 拉取\n\n```python\nollama.pull('gemma3')\n```\n\n### 推送\n\n```python\nollama.push('user\u002Fgemma3')\n```\n\n### 嵌入\n\n```python\nollama.embed(model='gemma3', input='天空之所以是蓝色的，是因为瑞利散射。')\n```\n\n### 批量嵌入\n\n```python\nollama.embed(model='gemma3', input=['天空之所以是蓝色的，是因为瑞利散射。', '草地之所以是绿色的，是因为叶绿素。'])\n```\n\n### 进程列表\n\n```python\nollama.ps()\n```\n\n## 错误处理\n\n如果请求返回错误状态码，或者在流式传输过程中检测到错误，将会抛出异常。\n\n```python\nmodel = '不存在的模型'\n\ntry:\n  ollama.chat(model)\nexcept ollama.ResponseError as e:\n  print('错误:', e.error)\n  if e.status_code == 404:\n    ollama.pull(model)\n```","# Ollama Python 快速上手指南\n\n## 环境准备\n\n在开始之前，请确保满足以下前置条件：\n\n1. **安装并运行 Ollama**\n   - 访问 [Ollama 官网](https:\u002F\u002Follama.com\u002Fdownload) 下载并安装对应系统的 Ollama 服务。\n   - 确保 Ollama 服务已在后台运行。\n\n2. **拉取模型**\n   - 在使用库之前，需先通过命令行拉取一个模型（例如 `gemma3`）：\n     ```sh\n     ollama pull gemma3\n     ```\n   - 更多可用模型请访问 [Ollama 模型搜索页](https:\u002F\u002Follama.com\u002Fsearch)。\n\n3. **Python 版本**\n   - 需要 Python 3.8 或更高版本。\n\n## 安装步骤\n\n使用 pip 安装官方 Python 库：\n\n```sh\npip install ollama\n```\n\n> **国内加速建议**：如果下载速度较慢，可使用国内镜像源安装：\n> ```sh\n> pip install ollama -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n> ```\n\n## 基本使用\n\n以下是最简单的对话示例，展示如何调用本地模型进行问答：\n\n```python\nfrom ollama import chat\n\nresponse = chat(model='gemma3', messages=[\n  {\n    'role': 'user',\n    'content': 'Why is the sky blue?',\n  },\n])\n\n# 打印回复内容\nprint(response['message']['content'])\n# 或者直接访问对象属性\nprint(response.message.content)\n```\n\n### 流式输出（可选）\n\n如果需要实时打印生成的内容（类似打字机效果），可开启流式模式：\n\n```python\nfrom ollama import chat\n\nstream = chat(\n    model='gemma3',\n    messages=[{'role': 'user', 'content': 'Why is the sky blue?'}],\n    stream=True,\n)\n\nfor chunk in stream:\n  print(chunk['message']['content'], end='', flush=True)\n```","某初创团队正在开发一款本地化智能客服系统，需要在 Python 后端中快速集成大语言模型以处理用户咨询，同时严格保障数据隐私不上传至第三方公有云。\n\n### 没有 ollama-python 时\n- **集成门槛高**：开发者需手动构建复杂的 HTTP 请求代码来调用 Ollama 本地服务，处理鉴权、超时重试及连接池管理耗费大量精力。\n- **流式响应难实现**：为了实现打字机效果的实时回复，必须自行解析 SSE（服务器发送事件）数据流，代码逻辑繁琐且容易出错。\n- **异步支持缺失**：在高并发场景下，同步阻塞式的请求方式会严重拖累服务器性能，而手动编写异步客户端不仅难度大且难以维护。\n- **类型安全无保障**：原生接口返回的是原始字典或 JSON 字符串，缺乏明确的类型提示，导致开发过程中频繁出现字段访问错误。\n\n### 使用 ollama-python 后\n- **极简集成体验**：仅需 `pip install` 并调用 `chat()` 函数即可连接本地模型，自动处理底层网络细节，将集成时间从数小时缩短至几分钟。\n- **原生流式支持**：通过设置 `stream=True` 直接获取生成器，轻松实现逐字输出的流畅交互，无需关心底层数据流解析逻辑。\n- **高效异步并发**：内置 `AsyncClient` 完美支持 `async\u002Fawait` 语法，让系统能轻松应对高并发请求，显著提升吞吐量而不增加代码复杂度。\n- **强类型开发辅助**：返回标准的 `ChatResponse` 对象，提供完整的 IDE 类型提示和自动补全，大幅减少运行时错误并提升代码可读性。\n\nollama-python 通过将复杂的模型调用封装为简洁的 Pythonic 接口，让开发者能专注于业务逻辑而非底层通信，极大地加速了本地大模型应用的落地进程。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Follama_ollama-python_ce1169ef.png","ollama","Ollama","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Follama_0cc73bd3.png","",null,"hello@ollama.com","https:\u002F\u002Follama.com","https:\u002F\u002Fgithub.com\u002Follama",[85],{"name":86,"color":87,"percentage":88},"Python","#3572A5",100,9724,995,"2026-04-07T01:49:08","MIT","未说明 (取决于 Ollama 后端支持的操作系统)","未说明 (本库为 Python 客户端，无本地 GPU 需求；实际推理依赖已安装的 Ollama 服务或云端 API)","未说明",{"notes":97,"python":98,"dependencies":99},"1. 必须先安装并运行 Ollama 服务端 (ollama.com\u002Fdownload)，本库仅作为客户端调用。2. 需预先通过命令行拉取模型 (如 ollama pull gemma3) 才能使用。3. 支持连接本地 Ollama 实例或 Ollama 云端 API (需配置 API Key)。4. 异步功能基于 asyncio。","3.8+",[100],"httpx",[15],[76,103],"python","2026-03-27T02:49:30.150509","2026-04-11T16:50:08.565122",[107,112,117,122,127,131],{"id":108,"question_zh":109,"answer_zh":110,"source_url":111},22530,"在使用工具（Tools）时，流式输出（Streaming）为什么不生效？","这是一个已知问题。当在 `chat` 函数中同时使用 `tools` 和 `stream=True` 时，响应不会增量返回，而是在所有工具调用完成后一次性返回。维护者已确认该问题将在下一个版本中修复。在此之前，如果需要使用工具，可能无法获得完美的流式体验。","https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fissues\u002F463",{"id":113,"question_zh":114,"answer_zh":115,"source_url":116},22531,"Ollama Python 库是否支持函数调用（Function Calling）或工具（Tools）？","是的，Ollama 现在支持将函数作为工具调用。官方已经发布了相关功能，您可以参考官方博客文章了解如何使用：\n1. 查看官方公告：https:\u002F\u002Follama.com\u002Fblog\u002Ffunctions-as-tools\n2. 查看工具支持详情：https:\u002F\u002Follama.com\u002Fblog\u002Ftool-support\n此外，社区也有处理多个工具同时调用的实践方案可供参考。","https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fissues\u002F39",{"id":118,"question_zh":119,"answer_zh":120,"source_url":121},22532,"遇到 `ollama._types.ResponseError` 错误且本地代理开启时如何解决？","该错误通常是因为开启了系统代理导致无法访问本地的 Ollama 服务（localhost）。\n解决方案：\n1. 临时关闭代理进行测试。\n2. 如果必须使用代理，可以通过配置 `Client` 来绕过代理或设置特定的请求头。例如：\n```python\nfrom ollama import Client\nclient = Client(\n  host='http:\u002F\u002Flocalhost:11434',\n  headers={'x-some-header': 'some-value'}\n)\nresponse = client.chat(model='llama3.2', messages=[{'role': 'user', 'content': 'Why is the sky blue?'}])\n```\n3. 更多关于在代理后使用 Ollama 的细节可参考官方 FAQ。","https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fissues\u002F396",{"id":123,"question_zh":124,"answer_zh":125,"source_url":126},22533,"在使用多模态模型（如 llama3.2-vision）时，设置 `raw=True` 后图片无法被识别怎么办？","当设置 `raw=True` 时，库不会自动处理图片占位符，导致模型看不到图片。解决方法是在提示词（prompt）中手动添加图片占位符。\n具体步骤：\n1. 在您的 prompt 字符串中插入 `\u003Cimage>` 标签，或者根据模型要求使用特定格式（如 `[img-0]\u003Cimage>`）。\n2. 示例代码逻辑：\n```python\n# 手动替换占位符\nprompt = prompt.replace(\"\u003C|image|>\", \"[img-0]\u003Cimage>\")\n# 然后调用 generate，确保 images 参数依然传递\nollama.generate(model=\"llama3.2-vision\", prompt=prompt, raw=True, images=[img_bytes])\n```\n这样即使在使用 `raw=True` 时，多模态模型也能正确接收并处理图像数据。","https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fissues\u002F319",{"id":128,"question_zh":129,"answer_zh":130,"source_url":111},22534,"如果模型有可用工具但未调用，流式响应是否会受到影响？","是的，这是一个相关的问题现象。即使模型最终没有调用任何工具，只要请求中定义了 `tools` 参数，流式响应也可能表现为一次性返回所有内容，而不是增量流式输出。这与“使用工具时流式不工作”的问题是同一根源，需等待官方后续版本修复。",{"id":132,"question_zh":133,"answer_zh":134,"source_url":116},22535,"如何在 Ollama 中模拟类似 OpenAI 的工具调用功能？","不需要手动模拟层，Ollama 原生已支持工具调用功能。您不再需要通过 `ollama.addtool` 等自定义方式来实现。请直接使用官方更新的 API 和文档，它们提供了原生的 `tools` 参数支持，允许模型在对话中自动决定何时调用函数。请参考官方文档 https:\u002F\u002Follama.com\u002Fblog\u002Ftool-support 获取最新的集成方法。",[136,141,146,151,156,161,166,171,176,181,186,191,196,201,206,211,216,221,226,231],{"id":137,"version":138,"summary_zh":139,"released_at":140},136237,"v0.6.1","## 变更内容\n* client\u002Ftypes：由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F601 中添加了对 logprobs 的支持\n\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.6.0...v0.6.1","2025-11-13T23:01:55",{"id":142,"version":143,"summary_zh":144,"released_at":145},136238,"v0.6.0","## 变更内容\n* 客户端：由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F578 中添加了网页搜索和网页爬取功能。\n* 客户端：由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F583 中实现了在初始化时加载 OLLAMA_API_KEY。\n* 客户端\u002Ftypes：由 @npardal 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F584 中更新了网页搜索和抓取 API。\n\n* 示例：由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F585 中添加了用于 web_search 和 web_crawl 的 mcp 服务器。\n* 示例：由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F588 中实现了 GPT 开源浏览器工具。\n\n## 新贡献者\n* @npardal 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F584 中完成了首次贡献。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.5.4...v0.6.0","2025-09-24T22:45:32",{"id":147,"version":148,"summary_zh":149,"released_at":150},136239,"v0.5.4","## 变更内容\n* 示例：@ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F558 中添加了 gpt-oss 浏览器示例\n* 构建（依赖项）：@dependabot[bot] 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F559 中将 actions\u002Fcheckout 从 4 升级到 5\n* 示例\u002Fgpt-oss：@ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F566 中修复了示例\n* 修复文档中 thinking-levels.py 的链接：@btjanaka 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F567 中完成\n* 示例：@ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F568 中修复了用于添加工具调用的 gpt-oss-tools-stream\n* 示例：解决 LLM 出错时 gpt-oss 工具使用状态码 400 的问题：@MarkWard0110 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F569 中完成\n* 构建（依赖项）：@dependabot[bot] 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F571 中将 actions\u002Fsetup-python 从 5 升级到 6\n* 功能：@mxyng 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F574 中为嵌入请求添加维度\n\n## 新贡献者\n* @btjanaka 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F567 中完成了首次贡献\n* @MarkWard0110 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F569 中完成了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.5.3...v0.5.4","2025-09-16T00:24:58",{"id":152,"version":153,"summary_zh":154,"released_at":155},136240,"v0.5.3","## 变更内容\n* @drifkin 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F553 中添加了对 ‘high’\u002F‘medium’\u002F‘low’ 思维值的支持\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.5.2...v0.5.3","2025-08-07T21:43:45",{"id":157,"version":158,"summary_zh":159,"released_at":160},136241,"v0.5.2","## 变更内容\n* 类型\u002F示例：@ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F537 中为消息和示例添加了 `tool_name` 字段。\n* 类型：@ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F538 中为 `ProcessResponse` 添加了 `context_length` 字段。\n* 类型：@ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F550 中放宽了工具类型的约束。\n\n* @ViViDboarder 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F526 中为包添加了许可证元数据。\n\n\n## 新贡献者\n* @hwittenborn 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F525 中完成了首次贡献。\n* @ViViDboarder 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F526 中完成了首次贡献。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.5.1...v0.5.2","2025-08-05T23:01:42",{"id":162,"version":163,"summary_zh":164,"released_at":165},136242,"v0.5.1","## 变更内容\n* 由 @drifkin 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F524 中为 `generate()` 完全添加了思维支持\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.5.0...v0.5.1","2025-05-30T21:32:27",{"id":167,"version":168,"summary_zh":169,"released_at":170},136243,"v0.5.0","## 变更内容\n* 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F521 中，由 @drifkin 添加了对思考功能的支持\n\n## 新贡献者\n* @drifkin 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F521 中完成了他们的首次贡献\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.9...v0.5.0","2025-05-30T02:38:51",{"id":172,"version":173,"summary_zh":174,"released_at":175},136244,"v0.4.9","## 变更内容\n* 使用 hatch 代替 poetry，由 @mxyng 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F445 中完成\n* 增加显示响应的功能，由 @jayrinaldime 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F511 中完成\n\n## 新贡献者\n* @jayrinaldime 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F511 中完成了首次贡献\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.8...v0.4.9","2025-05-27T18:08:58",{"id":177,"version":178,"summary_zh":179,"released_at":180},136245,"v0.4.8","## 变更内容\n* client：添加支持传入 Image 类型以进行生成，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F408 中实现\n* types：放宽工具的枚举类型，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F498 中实现\n* types：允许工具属性为单个或多个类型，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F499 中实现\n* types：启用传递具有任意角色的消息，由 @rylativity 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F462 中实现\n* types：允许工具使用 items 和 defs，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F501 中实现\n* examples：更新 tools.py，以优雅地处理数字而非拼接字符串，由 @ddbit 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F443 中实现\n* examples：修复示例链接，由 @ansent788 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F461 中实现\n* chore：添加 SECURITY.md 文件，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F479 中实现\n* poetry：更新依赖项，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F505 中实现\n\n\n## 新贡献者\n* @ddbit 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F443 中完成了首次贡献\n* @rylativity 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F462 中完成了首次贡献\n* @ansent788 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F461 中完成了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.7...v0.4.8","2025-04-16T21:54:38",{"id":182,"version":183,"summary_zh":184,"released_at":185},136246,"v0.4.7","## 变更内容\n* 客户端：移除 sha256，因为它已包含在摘要中，由 @samhita-alla 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F420 中完成。\n* 客户端：改进连接失败时的错误提示信息，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F398 中完成。\n* 杂项：将 httpx 从 0.27.2 升级到 0.28.1，由 @dependabot 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F365 中完成。\n\n* 杂项：确保导入语句按顺序排列，由 @akx 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F385 中完成。\n* 示例：修复创建示例，由 @jmorganca 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F421 中完成。\n* 示例：更新创建示例，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F418 中完成。\n* 修复：不在所有 Python 版本上运行代码风格检查，由 @akx 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F400 中完成。\n* 修复：格式化与 CI，由 @akx 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F422 中完成。\n* 杂项：为 blob 请求中的 sha 校验添加测试，由 @ParthSareen 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F425 中完成。\n\n## 新贡献者\n* @samhita-alla 在 https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F420 中完成了首次贡献。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.6...v0.4.7","2025-01-21T18:51:14",{"id":187,"version":188,"summary_zh":189,"released_at":190},136247,"v0.4.6","## What's Changed\r\n* add create api changes to support ollama v0.5.5 by @pdevine in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F417\r\n\r\n## New Contributors\r\n* @pdevine made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F417\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.5...v0.4.6","2025-01-14T01:31:39",{"id":192,"version":193,"summary_zh":194,"released_at":195},136248,"v0.4.5","## What's Changed\r\n* examples: remove extra print by @ParthSareen in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F364\r\n* examples: use type hinting generics in standard collections. by @fujitatomoya in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F389\r\n* bugfix: export Image model by @akx in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F384\r\n* bugfix: fix passing Image type in messages for chat @ParthSareen in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F390\r\n\r\n## New Contributors\r\n* @akx made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F384\r\n* @fujitatomoya made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F389\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.4...v0.4.5","2024-12-29T23:11:04",{"id":197,"version":198,"summary_zh":199,"released_at":200},136249,"v0.4.4","## What's Changed\r\n* fix validation of `format` to allow empty strings as it did previously by @jmorganca in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F369\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.3...v0.4.4","2024-12-08T03:38:42",{"id":202,"version":203,"summary_zh":204,"released_at":205},136250,"v0.4.3","## What's Changed\r\n* Improve tool example to showcase chatting by @ParthSareen in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F352\r\n* Bump pytest from 8.3.3 to 8.3.4 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F358\r\n* Structured outputs support with examples by @ParthSareen in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F354\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.2...v0.4.3","2024-12-06T20:04:12",{"id":207,"version":208,"summary_zh":209,"released_at":210},136251,"v0.4.2","## What's Changed\r\n* Fix message append in example chat-with-history.py to overcome runtime error. by @shahil-yadav in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F337\r\n* Make subscriptable methods more consistent with `__contains__` by @jmorganca in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F336\r\n* Fix `Client.chat` overloaded type annotations by @pythongirl325 in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F344\r\n* Fix image serialization for long image string by @ParthSareen in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F348\r\n\r\n## New Contributors\r\n* @shahil-yadav made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F337\r\n* @pythongirl325 made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F344\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.1...v0.4.2","2024-11-28T22:15:28",{"id":212,"version":213,"summary_zh":214,"released_at":215},136252,"v0.4.1","## What's Changed\r\n* Fixed issue where message types would have a `tool_calls` field by membership test even if no tool calls were provided by the response.\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.4.0...v0.4.1","2024-11-24T02:30:40",{"id":217,"version":218,"summary_zh":219,"released_at":220},136253,"v0.4.0","## New Features\r\n* Add Pydantic for validation and serialization by @mxyng in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F276\r\n* Passing Functions as Tools by @ParthSareen in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F321\r\n* No head request for create blob by @mxyng in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F277\r\n* Deprecate embeddings from docs in favor of embed by @royjhan in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F216\r\n\r\n## Chores\r\n* Examples refactor by @ParthSareen in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F329\r\n* chore: add python3.13 to test matrix by @mxyng in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F324\r\n* Update requirements.txt and poetry.lock by @ParthSareen in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F315\r\n* update pyproject.toml by @mxyng in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F284\r\n* Bump pytest from 8.3.2 to 8.3.3 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F278\r\n* Bump ruff from 0.6.3 to 0.6.5 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F281\r\n* Bump pydantic from 2.9.0 to 2.9.2 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F318\r\n* Bump ruff from 0.6.9 to 0.7.4 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F322\r\n* add basic delete\u002Fcopy tests by @mxyng in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F275\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.3.3...v0.4.0","2024-11-21T23:22:27",{"id":222,"version":223,"summary_zh":224,"released_at":225},136254,"v0.3.3","## What's Changed\r\n* _stream in async client raises RuntimeError when processing HTTP errors by @Oneirag in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F266\r\n* Bump ruff from 0.6.2 to 0.6.3 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F267\r\n\r\n## New Contributors\r\n* @Oneirag made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F266\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.3.2...v0.3.3","2024-09-09T17:22:54",{"id":227,"version":228,"summary_zh":229,"released_at":230},136255,"v0.3.2","## What's Changed\r\n* Add URL path to client URL in in Client._parse_host() by @bplunkert in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F170\r\n* IPv6 support by @jbinder in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F262\r\n* Bump pytest-httpserver from 1.0.12 to 1.1.0 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F252\r\n* Bump pytest-asyncio from 0.23.8 to 0.24.0 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F260\r\n* Bump ruff from 0.5.5 to 0.6.2 by @dependabot in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F261\r\n\r\n## New Contributors\r\n* @bplunkert made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F170\r\n* @jbinder made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F262\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.3.1...v0.3.2","2024-08-27T23:57:30",{"id":232,"version":233,"summary_zh":234,"released_at":235},136256,"v0.3.1","## What's Changed\r\n* update to `llama3.1` by @jmorganca in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F237\r\n* fix: tool_calls response parsing by @zeelrupapara in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F236\r\n* Update `Message` annotations to support tool calls by @Shulyaka in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F227\r\n\r\n## New Contributors\r\n* @jmorganca made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F237\r\n* @zeelrupapara made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F236\r\n* @Shulyaka made their first contribution in https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fpull\u002F227\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Follama\u002Follama-python\u002Fcompare\u002Fv0.3.0...v0.3.1","2024-07-29T17:41:37"]