免费资源 | 计算机基础教程

# 免费资源

服务器:

kaggle
google colab
www.oracle.com/cloud/free/ (opens new window)

awesome-free-llm-apis：把永久免费 LLM API 先整理成一张可直接拿来用的清单 (opens new window) A list of free LLM inference resources accessible via API. (opens new window)

几百个免费 AI 模型配额，一键接入本地项目。 (opens new window)

Agnes-AI 全模态 (opens new window)
Mistral AI 每月 10 亿 Token 免费，无需信用卡 Provider: Custom / OpenAI-compatible Base URL: https://api.mistral.ai/v1 Model: mistral-small-latest 或 mistral-medium-latest API Key: 从 console.mistral.ai 获取
Groq (opens new window)
- Llama 3.3 70B（当年最强开源模型之一）
- Mixtral 8x7B
- Gemma 2 9B
Google Gemini (opens new window)
- Gemini 3 Pro
- Gemini 3 Flash
- gemini-3.1-flash-lite-preview
GitHub Models
OpenRouter (opens new window)
- openrouter/free
- Nemotron 3 Super（free）
- Hunter Alpha
- Healer Alpha 多模态支持图文音理解
Opencode (opens new window)
- minimax-m2.5-free
- big-pickle
- GPT 5 Nano
cloudflare ai worker
- how to use
  1. run via rest api (opens new window)
  2. run via worker (opens new window)
- models (opens new window)
  1. @cf/runwayml/stable-diffusion-v1-5-inpainting (opens new window)
https://groq.com/
huggingface.co
英伟达nvidia (opens new window)
- Kimi-K2.5
- GLM-4.7
- z-ai/glm4.7
- minimaxai/minimax-m2.1
Cerebras
modelscope ModelScope provides a highly generous free API inference tier, offering a daily combined quota of 2,000 free API calls across hundreds of popular models (including Qwen, DeepSeek, GLM, and MiniMax)
硅基流动API官网：https://cloud.siliconflow.cn/i/jPI17VGX
- Kwai-Kolors/Kolors
- Qwen/Qwen3.5-4B
火山方舟API官网：https://console.volcengine.com
阿里百炼API官网：https://www.aliyun.com/product/bailian
- qwen3-max
深度求索API官网：https://platform.deepseek.com
智谱 AI
- GLM-4.7-Flash
月之暗面 Kimi
国家超算互联网

# Info

A Simple and Universal Swarm Intelligence Engine, Predicting Anything. 简洁通用的群体智能引擎，预测万物 (opens new window)

微舆：人人可用的多Agent舆情分析助手，打破信息茧房，还原舆情原貌，预测未来走向，辅助决策！从0实现，不依赖任何框架。 (opens new window)

简单的舆情监控分析 - 多平台热点聚合+基于 MCP 的AI分析工具。监控35个平台（抖音、知乎、B站、华尔街见闻、财联社等），智能筛选+自动推送+AI对话分析（用自然语言深度挖掘新闻：趋势追踪、情感分析、相似检索等13种工具）。 (opens new window)

Real-time global intelligence dashboard — AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface (opens new window)

Real-time dashboard for monitoring global news, markets, and geopolitical events (opens new window)

# todo

AI agents running research on single-GPU nanochat training automatically (opens new window)

笔记本跑LLM Official inference framework for 1-bit LLMs (opens new window)

Segment Anything Model (SAM) https://mp.weixin.qq.com/s/-LWG3rOz60VWiwdYG3iaWQ

Winglang，一门AI时代的全新编程语言 https://mp.weixin.qq.com/s/43k6Flpj5RhfT3MOxbF4aQ

SpatialLM: Training Large Language Models for Structured Indoor Modeling (opens new window)

Social Analyzer (opens new window)

HexStrike AI MCP Agents is an advanced MCP server that lets AI agents (Claude, GPT, Copilot, etc.) autonomously run 150+ cybersecurity tools for automated pentesting, vulnerability discovery, bug bounty automation, and security research. Seamlessly bridge LLMs with real-world offensive security capabilities. (opens new window)

excel 手搓 transformer (opens new window)

Implement a ChatGPT-like LLM in PyTorch from scratch, step by step by sebastian raschka (opens new window)

AI模型训练到底在训练什么？ (opens new window)

图解深度学习 - 数据蒸馏和知识蒸馏 Data Distillation (opens new window)

Model Distillation 终于把神经网络中的知识蒸馏搞懂了！！ (opens new window)

Transformer Explained Visually: Learn How LLM Transformer Models Work with Interactive Visualization (opens new window)

终于有人将深度学习中重点做成了动画 (opens new window)

GPT4All: Run Local LLMs on Any Device. Open-source and available for commercial use. (opens new window)

一饭封神：一个基于 AI 的智能菜谱生成平台，支持中华八大菜系 + 国际料理，提供营养分析、酒水推荐、菜谱效果图生成等全方位烹饪指导。 (opens new window)

AI首席情报官 Wiseflow (opens new window)

保姆级教程~本地微调DeepSeek-R1-8b模型 (opens new window)

Axolotl微调Qwen2-7b开源大模型！AutoGen+Marker自动生成数据集！AutoGen Studio+GPT4评估Qwen2文章能力 (opens new window)

rtvi-ai + pipecat https://github.com/rtvi-ai https://github.com/pipecat-ai/pipecat

我让AI当黑客！基于DeepSeek+ollama的网安神器开发实录，效率提升300% (opens new window)

Aider + Claude + Invoke + Udio ：从头到尾使用 AI 创建游戏！（音乐、代码、动画） (opens new window)

LLM在金融交易领域的实际应用:FinRobot开源平台的创新突破 (opens new window)

Restoration for TEMPEST images using deep-learning (opens new window)

多模态 multimodal Mini-Omni: Language Models Can Hear, Talk While Thinking in Streaming (opens new window)

API脚本

API管理分发
- LLM API 管理 & 分发系统 (opens new window)
思考过程脚本：
- 显示思考过程转换项目 (opens new window) 源自：chadyi 大佬的显示思考+搜索项目 (opens new window)

openwebui (opens new window)

// Before proceeding, ensure you're using Python 3.11 to avoid compatibility issues.
pip install open-webui
open-webui serve //Running on  http://0.0.0.0:8080

Setting:

Enable Web Search Engine
Connections
- OpenAI API
- Ollama API
  - http://localhost:11434

desktop

cherry-ai (opens new window) 没有联网搜索功能，不过其中小程序如秘塔搜索使用了满血的deepseek可以补足

brower Extension

Page Assist

AI+RAG AnythingLLM The all-in-one Desktop & Docker AI application with built-in RAG, AI agents (opens new window)

# 基础

# 私人部署

推理框架允许开发人员轻松地在本地部署和运行 LLM
- Ollama
- vllm
AI 应用开发平台提供了一套完整的工具和 API，用于构建、管理和部署 AI 应用
- Dify (opens new window)

https://mp.weixin.qq.com/s/iwsSKjGloNe7SCM4-5nzDA

# NextJS+Vercel

https://sdk.vercel.ai/

https://www.youtube.com/watch?v=5_JT9H8j1Us https://github.com/a-bit-of-saas/examples/tree/main/next/chat-gpt?ref=ethanmick.com

# FastGPT

代理方案：

nginx 方案 - 需要一台国外服务器 (商用)
sealos nginx 方案 -（约等于不用钱，不需要额外准备任何东西） labring/FastGPT/blob/main/docs/deploy/proxy/sealos.md
vercel反向代理（完全免费，github,google等，万能代理，免费100G流量） gaboolic/vercel-reverse-proxy
cloudflare 方案 - 需要有域名（每日免费 10w 次代理请求） zhuanlan.zhihu.com/p/622765456
clash 方案 - 仅需一台服务器（需要有 clash） labring/FastGPT/blob/main/docs/deploy/proxy/clash.md

https://github.com/labring/FastGPT 支持 sealos 、one-api（one-api:OpenAI 接口管理 & 分发系统，支持多种渠道包括 Azure，可用于二次分发管理 key， Docker 镜像，一键部署，开箱即用 https://github.com/songquanpeng/one-api）

github.com/Yidadaa/ChatGPT-Next-Web (opens new window) 支持vercel免费一键部署

# NAS gpt

https://mariushosting.com/how-to-install-chat-with-gpt-on-your-synology-nas/

https://cloud.tencent.com/developer/article/2312981

# 个人知识库 local rag

chatpgt GPTs
数据库助手 https://github.com/panosperi/superduperdb

# 去中心 web3+AI

https://github.com/OrbisWeb3/orbis-chat-gpt

https://www.gizatech.xyz/ https://www.youtube.com/watch?v=BHdP5oSISQE

# 自动化

Drag & drop UI to build your customized LLM flow https://github.com/FlowiseAI/Flowise

# community

开源 huggingface https://huggingface.co/

https://latentbox.com/en

# 原理学习

https://github.com/naklecha/llama3-from-scratch

https://github.com/karpathy/LLM101n

可视化 https://github.com/bbycroft/llm-viz

# 按模型分

# LLM - Large Language Models

开源
- LangChain
- LlamaIndex
- 手机上运行 https://h2o.ai/platform/danube/
闭源
- chatgpt
- claude

# Large Vision Models (LVM)

# Text to image model

Stable Diffusion model Stream Diffusion https://github.com/cumulo-autumn/StreamDiffusion https://colab.research.google.com/github/hewis123/stream-d/blob/main/Untitled2.ipynb

# ROBOT

Dobb·E An open-source, general framework for learning household robotic manipulation (opens new window)

# 按平台分类

ChatGPT

GOOGLE Bard

Cloudflare Workers AI https://mp.weixin.qq.com/s/R1m7IMUfHq53yEe-b9cjlA

# 按应用分类

# BOT

https://www.coze.com/ 使用扣子Coze——你想给你的抖音搭建一个24小时营业的机器人吗？ https://mp.weixin.qq.com/s/GZXfoTRROP15ZKf8GP0cug

# 图像增强

Demofusion

# 图片生成

Dalle3
Bing上的GPT-4V 完全免费使用gpt-4最新的多模态和最强图片生成模型 https://mp.weixin.qq.com/s/MNIZCA9pxyXobA3VVEOEqQ

logo制作： https://www.uugai.com/

https://logodiffusion.com/

# 标题/文案

https://duozhongcao.com/ https://www.jasper.ai/

https://hix.ai/

https://mymap.ai/

# PPT

MotionGo是必优科技(原口袋动画团队)全新升级的一款PPT动画插件,兼容WPS和office软件,轻量级产品,让PPT动效表达更专业。 https://motion.yoo-ai.com/?viaurl=ainavpro.com

https://www.beautiful.ai/

https://gamma.app/?lng=en

# 2D => 3D, 3D建模

https://convert.leiapix.com/ https://lumalabs.ai/interactive-scenes

# 市场分析

transform your product knowledge into a detailed Ideal Customer Profile https://m1-project.com

# 家居设计

https://collov.ai/