{"id":80541,"url":"https://github.com/xiaomingx/awesome-llm","name":"awesome-llm","description":"Awesome-LLM: a curated list of Large Language Model.🔥 大型语言模型（LLM）已经席卷了 全球，不再局限于 NLP 或 AI 社区。这里整理了一些关于大型语言模型，特别是与 ChatGPT 相关的研究论文，涵盖了 LLM 训练框架、部署工具、课程与教程，以及所有公开的 LLM 检查点和 API。","projects_count":99,"last_synced_at":"2026-09-16T00:00:43.730Z","repository":{"id":266043645,"uuid":"897202292","full_name":"XiaomingX/awesome-llm","owner":"XiaomingX","description":"Awesome-LLM: a curated list of Large Language Model.🔥 大型语言模型（LLM）已经席卷了 全球，不再局限于 NLP 或 AI 社区。这里整理了一些关于大型语言模型，特别是与 ChatGPT 相关的研究论文，涵盖了 LLM 训练框架、部署工具、课程与教程，以及所有公开的 LLM 检查点和 API。","archived":false,"fork":false,"pushed_at":"2026-02-03T12:33:45.000Z","size":148,"stargazers_count":67,"open_issues_count":2,"forks_count":13,"subscribers_count":3,"default_branch":"main","last_synced_at":"2026-08-07T20:22:05.866Z","etag":null,"topics":["agent","awesome","awesome-list","awesome-lists","gpt","llm","python","pytorch","tutorial"],"latest_commit_sha":null,"homepage":"https://x.com/seclink","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/XiaomingX.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-12-02T08:10:58.000Z","updated_at":"2026-07-30T13:42:33.000Z","dependencies_parsed_at":"2024-12-02T09:25:53.666Z","dependency_job_id":"5e3a2790-0476-49d0-b9d4-b9829be72dd8","html_url":"https://github.com/XiaomingX/awesome-llm","commit_stats":null,"previous_names":["xiaomingx/awesome-llm"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/XiaomingX/awesome-llm","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/XiaomingX%2Fawesome-llm","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/XiaomingX%2Fawesome-llm/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/XiaomingX%2Fawesome-llm/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/XiaomingX%2Fawesome-llm/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/XiaomingX","download_url":"https://codeload.github.com/XiaomingX/awesome-llm/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/XiaomingX%2Fawesome-llm/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":36934137,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-08-22T15:14:58.755Z","status":"online","status_checked_at":"2026-08-27T02:00:07.166Z","response_time":96,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"created_at":"2024-12-02T15:13:19.779Z","updated_at":"2026-09-16T00:00:43.730Z","primary_language":"Python","list_of_lists":false,"displayable":true,"categories":["重要论文里程碑","LLM应用","大型语言模型（LLM）排行榜","LLM教程与课程 (2025版)","LLM相关思考","LLM教程与课程","其他资源","LLM部署与本地运行","LLM书籍推荐","LLM部署","LLM 部署与推理 (Deployment \u0026 Inference)","热门 LLM 项目","其他相关论文","多模态预训练 (Multimodal Pre-training)","LLM训练框架","提示工程与优化 (Prompt Engineering)","LLM应用与智能体 (Agent)","强化学习与机器人 (RL \u0026 Robotics)","多模态嵌入 (Multimodal Embeddings)","LLM 编程助手 (AI Coding Assistants)","LLM 预训练方法 (Pre-training Methods)","LLM 加速与内核 (Acceleration \u0026 Kernels)","模型仓库与数据管理 (Model Registries)"],"sub_categories":["LLM 评估工具","开源大语言模型 (LLM)","LLM 评估与数据","推理网关 (Inference Gateways)","语音 (Audio)","强化学习 (Reinforcement Learning)","视频 (Video)","具身智能 (Robotics/Embodied AI)","推理引擎 (Inference Engines)","LLM 数据"],"readme":"# Awesome LLM (大型语言模型)\n\n🔥 大型语言模型（LLM）持续引领全球 AI 浪潮，已从通用的对话模型进化为具备深度推理（Reasoning）和自主行动（Agentic）能力的智能体。这里整理了关于大型语言模型，特别是与 ChatGPT、Claude、DeepSeek 及 Llama 系列相关的研究论文，涵盖了 LLM 训练框架、部署工具、课程与教程，以及所有公开的 SOTA 模型检查点和 API。\n\n## 热门 LLM 项目\n\n* **[DeepSeek-R1](https://github.com/deepseek-ai/DeepSeek-R1)** - 开源界的推理模型里程碑，通过强化学习激发了类似 OpenAI o1 的深度思考能力。\n* **[OpenAI o1/o3](https://openai.com/o1)** - 开启了“系统2”思维时代，通过思维链（CoT）在复杂数学、编程和科学问题上实现了突破性表现。\n* **[Llama 4 / 3.3](https://llama.meta.com/)** - Meta 继续引领开源，最新的 Llama 4 系列在多模态理解和长上下文处理上设立了新标准。\n* **[Gemini 3](https://deepmind.google/technologies/gemini/)** - Google 的多模态跃进之作，支持百万级 Token 上下文，具备强大的 UI 生成及前端编码能力。\n* **[Ollama](https://github.com/ollama/ollama)** - 目前最流行的本地 LLM 运行工具，支持一键运行 Llama 3.3, Gemma 2, Mistral 等模型。\n\n## 2026 发展趋势 (Trends)\n\n\u003e 💡 **核心洞察**: 大语言模型正从单一语言处理向**世界模型 (World Models)** 演进，深度融入机器感知、内容生成及因果推理能力。\n\n* **技术演进**: 2026年主流机构（如 DeepSeek, OpenAI, Apple）将密集迭代模型，重点攻克长程推理和 Agentic Workflow。\n* **企业策略**: 建议科技企业关注 Llama/Qwen 等开源方案用于内部研发，同时追踪高性价比 API 以优化运营成本。\n\n\n## 重要论文里程碑\n\n| 日期 | 关键词 | 机构 | 论文 |\n| --- | --- | --- | --- |\n| 2017-06 | Transformers | Google | [Attention Is All You Need](https://arxiv.org/pdf/1706.03762.pdf) |\n| 2018-06 | GPT 1.0 | OpenAI | [Improving Language Understanding by Generative Pre-Training](https://www.cs.ubc.ca/~amuham01/LING530/papers/radford2018improving.pdf) |\n| 2018-10 | BERT | Google | [BERT: Pre-training of Deep Bidirectional Transformers...](https://aclanthology.org/N19-1423.pdf) |\n| 2019-02 | GPT 2.0 | OpenAI | [Language Models are Unsupervised Multitask Learners](https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf) |\n| 2020-05 | GPT 3.0 | OpenAI | [Language models are few-shot learners](https://papers.nips.cc/paper/2020/file/1457c0d6bfcb4967418bfb8ac142f64a-Paper.pdf) |\n| 2022-01 | COT | Google | [Chain-of-Thought Prompting Elicits Reasoning in LLMs](https://arxiv.org/pdf/2201.11903.pdf) |\n| 2022-11 | ChatGPT | OpenAI | [Introducing ChatGPT](https://openai.com/blog/chatgpt) |\n| 2023-02 | LLaMA | Meta | [LLaMA: Open and Efficient Foundation Language Models](https://research.facebook.com/publications/llama-open-and-efficient-foundation-language-models/) |\n| 2023-03 | GPT-4 | OpenAI | [GPT-4 Technical Report](https://arxiv.org/abs/2303.08774) |\n| 2023-12 | Mamba | CMU/Princeton | [Mamba: Linear-Time Sequence Modeling with Selective State Spaces](https://arxiv.org/pdf/2312.00752) |\n| 2024-01 | DeepSeek-MoE | DeepSeek | [DeepSeekMoE: Towards Ultimate Expert Specialization...](https://arxiv.org/abs/2401.06066) |\n| 2024-05 | Mamba2 | CMU/Princeton | [Transformers are SSMs: Generalized Models...](https://arxiv.org/abs/2405.21060) |\n| 2024-06 | Claude 3.5 | Anthropic | [Claude 3.5 Sonnet System Card](https://www.anthropic.com/research) |\n| 2024-07 | Llama 3.1 | Meta | [The Llama 3 Herd of Models](https://arxiv.org/abs/2407.21783) |\n| 2024-09 | OpenAI o1 | OpenAI | [Learning to Reason with LLMs (System Card)](https://openai.com/index/learning-to-reason-with-llms/) |\n| 2024-09 | Qwen 2.5 | Alibaba | [Qwen2.5 Technical Report](https://qwenlm.github.io/blog/qwen2.5/) |\n| 2024-12 | DeepSeek-V3 | DeepSeek | [DeepSeek-V3 Technical Report](https://arxiv.org/abs/2412.19437) |\n| 2025-01 | DeepSeek-R1 | DeepSeek | [DeepSeek-R1: Incentivizing Reasoning via Reinforcement Learning](https://github.com/deepseek-ai/DeepSeek-R1) |\n| 2025-05 | Claude 4 | Anthropic | [Claude 4 Opus Technical Report](https://www.anthropic.com/research) |\n| 2025-08 | GPT-5.2 | OpenAI | [GPT-5 System Card](https://openai.com/gpt-5) (专业工作优化，强化工具调用) |\n| 2025-09 | Claude 4.5 | Anthropic | [Claude 4.5 Sonnet Update](https://www.anthropic.com/news) |\n\n## 其他相关论文\n\n\n\n* **[LLM推理与思维链 (Reasoning \u0026 CoT)](https://github.com/Timothyxxx/Chain-of-ThoughtsPapers)** - 关注 OpenAI o1 和 DeepSeek-R1 引发的推理能力研究。\n* **[高效微调与量化](https://github.com/unslothai/unsloth)** - Unsloth, QLoRA 等让个人显卡也能微调大模型的技术。\n* **[Agentic Workflow (智能体工作流)](https://github.com/langchain-ai/langgraph)** - 从单一 Prompt 到复杂的智能体协作系统。\n\n## 大型语言模型（LLM）排行榜\n\n* **Chatbot Arena (LMSYS)** - [排行榜链接](https://www.google.com/search?q=https://chat.lmsys.org/) - 业内最权威的盲测排行榜，涵盖 GPT-5, Claude 4.5, Llama 4 等最新模型。\n* **OpenCompass 2.0** - [OpenCompass](https://rank.opencompass.org.cn) - 涵盖代码、数学、推理等维度的全方位评测。\n* **LiveBench** - [LiveBench](https://livebench.ai/) - 旨在防止“刷榜”的动态基准测试，题目定期更新，更能反映模型真实能力。\n* **Open LLM Leaderboard** - [Hugging Face](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) - 针对开源模型的权威自动化评测。\n\n### 开源大语言模型 (LLM)\n\n#### Meta\n\n* **[Llama 4 系列](https://llama.meta.com/)** - 最新的多模态开源王者 (Scout, Maverick, Behemoth)。\n* **[Llama 3.3 / 3.1](https://huggingface.co/meta-llama)** - 依然强大的工业级标准模型 (70B, 405B)。\n\n#### DeepSeek (深度求索)\n\n* **[DeepSeek-R1](https://github.com/deepseek-ai/DeepSeek-R1)** - 开源推理模型首选，数学与代码能力媲美闭源 SOTA。\n* **[DeepSeek-V3](https://huggingface.co/deepseek-ai)** - 极具性价比的 MoE 模型，性能对标 GPT-4o。\n\n#### Alibaba (阿里云)\n\n* **[Qwen 3 (通义千问)](https://github.com/QwenLM/Qwen)** - Qwen 3 系列在编码和数学领域持续霸榜。\n* **[Qwen 2.5](https://huggingface.co/Qwen)** - 覆盖 0.5B 到 72B 的全尺寸高性能模型。\n\n#### Mistral AI\n\n* **[Mistral Large 2](https://mistral.ai/)** - 欧洲最强开源模型，上下文窗口大，指令遵循能力强。\n* **[Mistral Nemo / Small](https://huggingface.co/mistralai)** - 针对端侧和低延迟场景优化。\n\n#### Google\n\n* **[Gemma 3 / 2](https://ai.google.dev/gemma)** - Google 的轻量级开源模型，适合学术研究和笔记本部署。\n\n#### 其他优秀模型\n\n* **[GLM-4-Plus](https://github.com/THUDM/GLM-4)** - 智谱AI 最新力作，结合 PPO 技术大幅提升推理与指令遵循能力，商业化落地首选。\n* **[Phi-4](https://huggingface.co/microsoft)** - Microsoft 推出的“小而美”模型，推理能力惊人。\n* **[Nemotron-4](https://huggingface.co/nvidia)** - NVIDIA 发布的用于生成合成数据的重型模型。\n\n### LLM 评估与数据\n\n* **[LiveBench](https://livebench.ai/)** - 防止数据泄露的动态评测。\n* **[Scale AI Leaderboard](https://scale.com/leaderboard)** - 专注于代码生成和指令遵循的评测。\n\n## LLM训练框架\n\n* **[Unsloth](https://github.com/unslothai/unsloth)** - **强烈推荐**。训练速度提升 2-5 倍，显存占用减少 50%，是目前微调 Llama/Mistral 的首选工具。\n* **[Axolotl](https://www.google.com/search?q=https://github.com/OpenAccess-AI-Collective/axolotl)** - 配置化微调工具，支持绝大多数主流开源模型。\n* **[Llama-Factory](https://github.com/hiyouga/LLaMA-Factory)** - 提供 WebUI 的微调框架，对中文用户非常友好。\n* **[Firefly](https://github.com/yangjianxin1/Firefly)** - 全能型训练框架，支持预训练、SFT、DPO，适配 Qwen/Llama 等主流模型。\n* **[DeepSpeed](https://github.com/microsoft/DeepSpeed)** \u0026 **[Megatron-LM](https://github.com/NVIDIA/Megatron-LM)** - 依然是超大规模集群预训练的基石。\n* **[ColossalAI](https://github.com/hpcaitech/ColossalAI)** - 面向大模型时代的统一并行训练系统，支持异构内存管理，显著降低显存开销。\n\n\u003e 💡 **选型建议**: 对于大规模预训练（千卡级），优先选择 **DeepSpeed** 或 **ColossalAI** 以确保稳定性和效率；对于中小规模微调或快速验证，**Unsloth** 和 **Llama-Factory** 是最高效的选择。\n\n## LLM 加速与内核 (Acceleration \u0026 Kernels)\n\n* **[FlashInfer](https://github.com/flashinfer-ai/flashinfer)** - 加速 FlashAttention，支持 2-5x 速度提升，LLM Serving 必备内核库。\n* **[DeepGEMM](https://github.com/deepseek-ai/DeepGEMM)** - DeepSeek 开源的高效 FP8 GEMM 内核，适用于极致性能优化。\n* **[DeepEP](https://github.com/deepseek-ai/DeepEP)** - 专家并行通信库，大幅提升 MoE 模型分布式训练效率。\n\n## 多模态预训练 (Multimodal Pre-training)\n\n### 语音 (Audio)\n* **[SpeechBrain](https://github.com/speechbrain/speechbrain)** - PyTorch 语音工具包，支持 wav2vec 2.0/Whisper 等大规模自监督预训练。\n* **[S3PRL](https://github.com/s3prl/s3prl)** - 专注于自监督学习 (SSL) 的预训练框架，支持 HuBERT/TERA 等 Upstream 模型。\n\n### 视频 (Video)\n* **[VideoMAE](https://github.com/MCG-NJU/VideoMAE)** - (V1/V2) 视频版 MAE，通过高掩码率自监督学习高效视频表示 (Kinetics-400 SOTA)。\n* **[VPT (Video Pre-Training)](https://github.com/openai/Video-Pre-Training)** - OpenAI 用行为克隆从 Minecraft 视频大规模预训练代理模型。\n* **[EVL](https://github.com/OpenGVLab/efficient-video-recognition)** - 冻结 CLIP 骨干的高效视频学习器，极低训练成本实现高精度识别。\n\n## LLM 预训练方法 (Pre-training Methods)\n\n* **[MiniMind](https://github.com/jingyaogong/minimind)** - 从零构建小参数 LLM (26M起) 的全流程教程，适合低成本实验。\n* **[SO-Large-LM](https://github.com/datawhalechina/so-large-lm)** - 系统化的预训练开源教程，覆盖数据清洗、分词到 MoE 架构设计。\n* **[Happy-LLM](https://github.com/datawhalechina/happy-llm)** - 专注 Transformer 原理实现的预训练实践项目。\n\n## 多模态嵌入 (Multimodal Embeddings)\n\n* **[ImageBind](https://github.com/facebookresearch/ImageBind)** - Meta 开源的六模态统一嵌入空间 (图像/文本/音频/深度/热/IMU)。\n* **[CLIP](https://github.com/openai/CLIP)** \u0026 **[timm](https://github.com/huggingface/pytorch-image-models)** - 图像文本对齐的基石模型与最大的视觉骨干库。\n* **[E5-V](https://github.com/kongds/E5-V)** - 桥接 Llama3 与视觉模型的通用多模态嵌入。\n\n## 强化学习与机器人 (RL \u0026 Robotics)\n\n### 强化学习 (Reinforcement Learning)\n* **[Stable-Baselines3](https://github.com/DLR-RM/stable-baselines3)** - 基于 PyTorch 的标准 RL 算法实现 (PPO, TD3, SAC)，稳定可靠。\n* **[RLlib (Ray)](https://github.com/ray-project/ray)** - 工业级分布式 RL 训练框架，支持大规模多智能体环境。\n* **[CleanRL](https://github.com/vwxyzjn/cleanrl)** - 单文件实现的 RL 算法库 (PPO/DQN)，代码极其简洁，适合入门与魔改。\n* **[Easy-RL](https://github.com/datawhalechina/easy-rl)** - 经典的强化学习中文教程（蘑菇书），配套完善的代码实践。\n\n### 具身智能 (Robotics/Embodied AI)\n* **[MuJoCo](https://github.com/google-deepmind/mujoco)** - DeepMind 开源的高精度物理引擎，RL 机器人研究的标准环境。\n* **[OpenManipulator](https://github.com/ROBOTIS-GIT/open_manipulator)** - 基于 ROS 的开源机械臂平台，支持完整的仿真到实物迁移。\n* **[RoboNet](https://github.com/google/roboc_suite)** - 大规模机器人交互数据集，用于泛化策略学习。\n\n\n## LLM 部署与推理 (Deployment \u0026 Inference)\n\n### 推理引擎 (Inference Engines)\n* **[vLLM](https://github.com/vllm-project/vllm)** - 生产环境首选。支持 PagedAttention 和连续批处理，吞吐量 SOTA。\n* **[llama.cpp](https://github.com/ggerganov/llama.cpp)** - 边缘计算神器。纯 C++ 实现，针对 Apple Silicon 和 CPU 极致优化。\n* **[TensorRT-LLM](https://github.com/NVIDIA/TensorRT-LLM)** - NVIDIA 官方加速库，支持 In-flight batching 和 FP8 量化。\n* **[LMDeploy](https://github.com/InternLM/lmdeploy)** - InternLM 团队推出的高性能推理工具，TurboMind 引擎速度极快。\n* **[LightLLM](https://github.com/ModelTC/lightllm)** - 纯 Python/Triton 实现的轻量级推理框架，易于修改和扩展。\n\n### 推理网关 (Inference Gateways)\n* **[Inference Gateway](https://github.com/inference-gateway/inference-gateway)** - 企业级 AI 网关，统一管理 OpenAI/Ollama/Anthropic 等接口，支持 MCP 协议。\n* **[LiteLLM](https://github.com/BerriAI/litellm)** - **最流行代理**。用统一的 OpenAI 格式调用 100+ 种 LLM API，支持负载均衡和成本追踪。\n* **[llm-d](https://github.com/llm-d/llm-d)** - Kubernetes 原生的分布式推理栈，集成了 vLLM 和网关。\n* **[Open WebUI](https://github.com/open-webui/open-webui)** - 功能最强大的本地 Web 界面，自带 RAG 和多模型管理。\n* **[BentoML](https://github.com/bentoml/BentoML)** - 将模型打包为生产级微服务的统一框架。\n\n## 模型仓库与数据管理 (Model Registries)\n\n* **[KohakuHub](https://github.com/KohakuBlueleaf/KohakuHub)** - 自托管的 Hugging Face 替代方案，支持 Git-like 版本控制，适合企业内部私有部署。\n* **[DagsHub](https://dagshub.com)** - 结合 GitHub + DVC 的开源 ML 平台，提供实验跟踪和数据版本管理。\n\n## LLM 编程助手 (AI Coding Assistants)\n\n* **[Cursor](https://www.cursor.com/)** - **当前体验最佳**。基于 VS Code 修改的 AI 原生编辑器，Tab 补全和 Composer 功能极大地改变了编程体验。\n* **[Windsurf](https://codeium.com/windsurf)** - Codeium 推出的 Agentic IDE，主打深度上下文感知和主动行动能力。\n* **[Cline](https://github.com/cline/cline)** - 开源的自主编程 Agent 插件，能够执行终端命令、文件读写，配合 Claude 3.5 Sonnet 效果拔群。\n* **[Continue](https://github.com/continuedev/continue)** - 开源的 IDE 插件 (VS Code / JetBrains)，支持连接本地 Ollama 模型进行辅助编程。\n\n## LLM应用与智能体 (Agent)\n\n* **[LangGraph](https://github.com/langchain-ai/langgraph)** - LangChain 的升级版，专注于构建有状态、循环的智能体工作流。\n* **[Dify](https://github.com/langgenius/dify)** - 开源的 LLM 应用开发平台，可视化编排 RAG 和 Agent。\n* **[MemGPT](https://github.com/cpacker/MemGPT)** - 赋予 LLM 长期记忆和操作系统级别的上下文管理能力。\n* **[CrewAI](https://github.com/joaomdmoura/crewAI)** - 编排角色扮演的 AI 智能体团队来共同完成任务。\n## 提示工程与优化 (Prompt Engineering)\n\n* **[DSPy](https://github.com/stanfordnlp/dspy)** - 斯坦福推出的框架，主张“编程”而非“提示”语言模型，通过编译自动优化 Prompt。\n* **[Promptfoo](https://github.com/promptfoo/promptfoo)** - 开发者友好的 LLM 测试工具，用于评估 Prompt 质量和模型输出，防止回归。\n\n## LLM教程与课程 (2025版)\n\n* **[Andrej Karpathy's LLM101n](https://github.com/karpathy/LLM101n)** - Karpathy 大神的最新课程，教你从零构建一个 Storyteller AI。\n* **[Generative AI with LLMs (Coursera)](https://www.coursera.org/learn/generative-ai-with-llms)** - AWS 和 DeepLearning.AI 联合推出的实战课程。\n* **[DeepLearning.AI Short Courses](https://www.deeplearning.ai/short-courses/)** - 包含 RAG、Agent、Fine-tuning 等大量免费短课程。\n","projects_url":"https://awesome.ecosyste.ms/api/v1/lists/xiaomingx%2Fawesome-llm/projects"}