#jupyter_notebook #gemini #large_language_models #llm #openai #training #transformers
This project shows how to build and train a transformer language model from scratch in PyTorch. It uses the Pile dataset, tokenizes text with tiktoken, and stores tokens in HDF5 files for faster training. The code includes attention, MLP, transformer blocks, training, saving, and text generation. The benefit is that you can learn how LLMs work and train your own small or large model on a single GPU, then use it to generate text for your own tasks.
https://github.com/FareedKhan-dev/train-llm-from-scratch
This project shows how to build and train a transformer language model from scratch in PyTorch. It uses the Pile dataset, tokenizes text with tiktoken, and stores tokens in HDF5 files for faster training. The code includes attention, MLP, transformer blocks, training, saving, and text generation. The benefit is that you can learn how LLMs work and train your own small or large model on a single GPU, then use it to generate text for your own tasks.
https://github.com/FareedKhan-dev/train-llm-from-scratch
GitHub
GitHub - FareedKhan-dev/train-llm-from-scratch: A straightforward method for training your LLM, from downloading data to generating…
A straightforward method for training your LLM, from downloading data to generating text. - FareedKhan-dev/train-llm-from-scratch
❤1
#go #hysteria #hysteria2 #naive_proxy #shadowsocks #shadowtls #sing_box #trojan #tuic #vless #vmess
S-UI is a web panel for managing SagerNet/Sing-Box. It supports many protocols, several languages, traffic routing, client and system status, and subscription links. It runs on Linux, Windows, and partly on macOS, and can be installed by script, manually, or with Docker. It also supports HTTPS and dark/light mode. This helps you set up and control your proxy service more easily from one place.
https://github.com/alireza0/s-ui
S-UI is a web panel for managing SagerNet/Sing-Box. It supports many protocols, several languages, traffic routing, client and system status, and subscription links. It runs on Linux, Windows, and partly on macOS, and can be installed by script, manually, or with Docker. It also supports HTTPS and dark/light mode. This helps you set up and control your proxy service more easily from one place.
https://github.com/alireza0/s-ui
GitHub
GitHub - alireza0/s-ui: An advanced Web Panel • Built for SagerNet/Sing-Box
An advanced Web Panel • Built for SagerNet/Sing-Box - alireza0/s-ui
#csharp #agent_skills
.NET Agent Skills is a set of curated skills and custom agents for coding tools that help with common .NET tasks like coding, data access, debugging, builds, NuGet, upgrades, MAUI, AI, templates, testing, ASP.NET, and new .NET 11 features. You can install them in Copilot CLI, Claude Code, VS Code, Cursor, or Codex CLI. The benefit is faster, more accurate .NET help with less guesswork, so you can build, fix, and improve projects more easily.
https://github.com/dotnet/skills
.NET Agent Skills is a set of curated skills and custom agents for coding tools that help with common .NET tasks like coding, data access, debugging, builds, NuGet, upgrades, MAUI, AI, templates, testing, ASP.NET, and new .NET 11 features. You can install them in Copilot CLI, Claude Code, VS Code, Cursor, or Codex CLI. The benefit is faster, more accurate .NET help with less guesswork, so you can build, fix, and improve projects more easily.
https://github.com/dotnet/skills
GitHub
GitHub - dotnet/skills: Repository for skills to assist AI coding agents with .NET and C#
Repository for skills to assist AI coding agents with .NET and C# - dotnet/skills
👍1
#python #agentic_ai #agentic_workflow #agents #function_calling #llama_cpp #llamafile #llm #ollama #python #self_hosted #tool_calling
Forge is a Python tool that makes self-hosted LLM tool-calling more reliable. It helps local models handle multi-step tasks with guardrails, better context control, and support for Ollama, llama-server, Llamafile, and Anthropic. You can use it as a workflow runner, middleware, or proxy server with OpenAI-style clients. The benefit is fewer broken tool calls, better results on small models, and easier setup for agent apps, chat tools, and long-running sessions.
https://github.com/antoinezambelli/forge
Forge is a Python tool that makes self-hosted LLM tool-calling more reliable. It helps local models handle multi-step tasks with guardrails, better context control, and support for Ollama, llama-server, Llamafile, and Anthropic. You can use it as a workflow runner, middleware, or proxy server with OpenAI-style clients. The benefit is fewer broken tool calls, better results on small models, and easier setup for agent apps, chat tools, and long-running sessions.
https://github.com/antoinezambelli/forge
GitHub
GitHub - antoinezambelli/forge: A Python framework for self-hosted LLM tool-calling and multi-step agentic workflows
A Python framework for self-hosted LLM tool-calling and multi-step agentic workflows - antoinezambelli/forge
#typescript #antigravity_skills #business_knowledge #claude_code #claude_skills #codebase_analysis #codex #codex_skills #developer_tools_ai_agent #gemini_cli_skills #karpathy_llm_wiki #knowledge_base #knowledge_graph #memory #opencode_skills #pi_agent #understandcode #vibe_coding
Understand Anything turns a codebase or docs into an interactive knowledge graph you can search, explore, and ask questions about. It works with tools like Claude Code, Codex, Cursor, Copilot, and Gemini CLI. You can see files, functions, classes, and dependencies in one place, get plain-English explanations, and find what changes affect before you commit. This helps you learn large projects faster, understand how pieces fit together, and save time when onboarding or reviewing code.
https://github.com/Lum1104/Understand-Anything
Understand Anything turns a codebase or docs into an interactive knowledge graph you can search, explore, and ask questions about. It works with tools like Claude Code, Codex, Cursor, Copilot, and Gemini CLI. You can see files, functions, classes, and dependencies in one place, get plain-English explanations, and find what changes affect before you commit. This helps you learn large projects faster, understand how pieces fit together, and save time when onboarding or reviewing code.
https://github.com/Lum1104/Understand-Anything
GitHub
GitHub - Lum1104/Understand-Anything: Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph…
Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, G...
❤1
#other
CFnew v2.9.8 is a Cloudflare Worker/Pages tool for managing VLESS, Trojan, and xhttp subscriptions with a simple web panel, custom paths, ECH support, delay testing, and many client formats. It now builds Clash, Sing-box, Surge, Loon, and Quantumult X configs directly, so setup is faster and more reliable. For you, this means easier deployment, instant config changes without redeploying, better performance, and more control over proxy and region settings.
https://github.com/byJoey/cfnew
CFnew v2.9.8 is a Cloudflare Worker/Pages tool for managing VLESS, Trojan, and xhttp subscriptions with a simple web panel, custom paths, ECH support, delay testing, and many client formats. It now builds Clash, Sing-box, Surge, Loon, and Quantumult X configs directly, so setup is faster and more reliable. For you, this means easier deployment, instant config changes without redeploying, better performance, and more control over proxy and region settings.
https://github.com/byJoey/cfnew
GitHub
GitHub - byJoey/cfnew
Contribute to byJoey/cfnew development by creating an account on GitHub.
#rust #analytics #bi #data_visualization #javascript #jupyter #python #real_time #webassembly
Perspective is a fast tool for exploring large and streaming data. It lets you build dashboards, reports, notebooks, and apps with tables and many chart types. It works in the browser, Python, and Rust, and can connect to data sources like DuckDB or Arrow. This helps you quickly see patterns, make better decisions, and analyze data without heavy setup.
https://github.com/perspective-dev/perspective
Perspective is a fast tool for exploring large and streaming data. It lets you build dashboards, reports, notebooks, and apps with tables and many chart types. It works in the browser, Python, and Rust, and can connect to data sources like DuckDB or Arrow. This helps you quickly see patterns, make better decisions, and analyze data without heavy setup.
https://github.com/perspective-dev/perspective
GitHub
GitHub - perspective-dev/perspective: A data visualization and analytics component, especially well-suited for large and/or streaming…
A data visualization and analytics component, especially well-suited for large and/or streaming datasets. - perspective-dev/perspective
👍2
#javascript #ai_agent #ai_presentation #api #gamma #powerpoint_automation #powerpoint_free #powerpoint_generation #presentation
Presenton is an open-source tool that makes AI slides and lets you use your own models, keys, templates, and data. You can run it in Docker, on your computer as a desktop app, or in the browser, and export slides as editable PPTX or PDF. This gives you more privacy, more control, and no lock-in or forced subscriptions.
https://github.com/presenton/presenton
Presenton is an open-source tool that makes AI slides and lets you use your own models, keys, templates, and data. You can run it in Docker, on your computer as a desktop app, or in the browser, and export slides as editable PPTX or PDF. This gives you more privacy, more control, and no lock-in or forced subscriptions.
https://github.com/presenton/presenton
GitHub
GitHub - presenton/presenton: Open-Source AI Presentation Generator and API (Gamma, Beautiful AI, Decktopus Alternative)
Open-Source AI Presentation Generator and API (Gamma, Beautiful AI, Decktopus Alternative) - presenton/presenton
👍1
#python #infra #long #nvfp4 #parallel #real_time #video_generation
LongLive 2.0 helps generate long videos faster by using NVFP4, parallel processing, and multi-shot support for training and inference. It can reach up to 45.7 FPS while keeping good quality, and it also supports real-time, user-guided video generation. The benefit is that you can make long videos more quickly and efficiently, with less computing cost and smoother interactive use.
https://github.com/NVlabs/LongLive
LongLive 2.0 helps generate long videos faster by using NVFP4, parallel processing, and multi-shot support for training and inference. It can reach up to 45.7 FPS while keeping good quality, and it also supports real-time, user-guided video generation. The benefit is that you can make long videos more quickly and efficiently, with less computing cost and smoother interactive use.
https://github.com/NVlabs/LongLive
GitHub
GitHub - NVlabs/LongLive: LongLive 2.0: Infra - Long Video Gen
LongLive 2.0: Infra - Long Video Gen. Contribute to NVlabs/LongLive development by creating an account on GitHub.
#python #ai_agents #claude_code #cloud_security #cybersecurity #devsecops #ethical_hacking #incident_response #infosec #llm #malware_analysis #mcp #mitre_attack #nist_csf #osint #penetration_testing #red_team #security #security_automation #threat_hunting #threat_intelligence
Anthropic Cybersecurity Skills is a free, open-source library with 754 ready-made security skills for AI agents across 26 domains. It helps an AI act more like a senior security analyst by giving clear workflows, checks, and verification steps. It also maps each skill to major security frameworks like MITRE ATT&CK, NIST CSF 2.0, ATLAS, D3FEND, and NIST AI RMF. The benefit is faster, smarter, and more reliable security work with less guesswork.
https://github.com/mukul975/Anthropic-Cybersecurity-Skills
Anthropic Cybersecurity Skills is a free, open-source library with 754 ready-made security skills for AI agents across 26 domains. It helps an AI act more like a senior security analyst by giving clear workflows, checks, and verification steps. It also maps each skill to major security frameworks like MITRE ATT&CK, NIST CSF 2.0, ATLAS, D3FEND, and NIST AI RMF. The benefit is faster, smarter, and more reliable security work with less guesswork.
https://github.com/mukul975/Anthropic-Cybersecurity-Skills
GitHub
GitHub - mukul975/Anthropic-Cybersecurity-Skills: 754 structured cybersecurity skills for AI agents · Mapped to 5 frameworks: MITRE…
754 structured cybersecurity skills for AI agents · Mapped to 5 frameworks: MITRE ATT&CK, NIST CSF 2.0, MITRE ATLAS, D3FEND & NIST AI RMF · agentskills.io standard · Works with Clau...
💩2🤬1
#python
Plugins help Claude act like a specialist for different jobs, such as sales, marketing, finance, data, and support. They connect Claude to your tools, add useful skills, and give quick slash commands for common tasks. You can also change them to match your company’s words, systems, and ways of working. This saves time, reduces repeat work, and gives more consistent results because Claude can work more like part of your team.
https://github.com/anthropics/knowledge-work-plugins
Plugins help Claude act like a specialist for different jobs, such as sales, marketing, finance, data, and support. They connect Claude to your tools, add useful skills, and give quick slash commands for common tasks. You can also change them to match your company’s words, systems, and ways of working. This saves time, reduces repeat work, and gives more consistent results because Claude can work more like part of your team.
https://github.com/anthropics/knowledge-work-plugins
GitHub
GitHub - anthropics/knowledge-work-plugins: Open source repository of plugins primarily intended for knowledge workers to use in…
Open source repository of plugins primarily intended for knowledge workers to use in Claude Cowork - anthropics/knowledge-work-plugins
#swift #amp #claude_code #codex #gemini #ghostty #opencode #terminal #tmux
cmux is a fast macOS terminal app made for AI coding work. It gives you vertical tabs, split panes, clear notifications, and a built-in browser, so you can manage many agent sessions in one place and quickly see which one needs attention. It also works with your Ghostty settings, supports SSH, custom commands, session restore, and automation through a CLI and API. This helps you stay organized, save time, and work more smoothly with coding agents.
https://github.com/manaflow-ai/cmux
cmux is a fast macOS terminal app made for AI coding work. It gives you vertical tabs, split panes, clear notifications, and a built-in browser, so you can manage many agent sessions in one place and quickly see which one needs attention. It also works with your Ghostty settings, supports SSH, custom commands, session restore, and automation through a CLI and API. This helps you stay organized, save time, and work more smoothly with coding agents.
https://github.com/manaflow-ai/cmux
GitHub
GitHub - manaflow-ai/cmux: Ghostty-based macOS terminal with vertical tabs and notifications for AI coding agents
Ghostty-based macOS terminal with vertical tabs and notifications for AI coding agents - manaflow-ai/cmux
#javascript #android #awesome #awesome_list #free_apps #ios #linux #macos #microsoft #mobile #windows
This page is a big list of free apps for Windows, Mac, Linux, Android, and iOS, sorted by what they do, like audio, browsers, security, video, games, editing, and file tools. It helps you quickly find a useful app for your device, especially if you want free or open-source software. The page also marks which apps are recommended and which are available on each platform, so you can choose faster and avoid wasted downloads.
https://github.com/Axorax/awesome-free-apps
This page is a big list of free apps for Windows, Mac, Linux, Android, and iOS, sorted by what they do, like audio, browsers, security, video, games, editing, and file tools. It helps you quickly find a useful app for your device, especially if you want free or open-source software. The page also marks which apps are recommended and which are available on each platform, so you can choose faster and avoid wasted downloads.
https://github.com/Axorax/awesome-free-apps
GitHub
GitHub - Axorax/awesome-free-apps: Curated list of the best free apps for PC and mobile
Curated list of the best free apps for PC and mobile - Axorax/awesome-free-apps
#other
This skill helps you remove AI-like writing patterns so your prose sounds more natural, direct, and human. It gives clear rules for cutting banned phrases, avoiding cliché structures, using active voice, and checking quality, which helps you write cleaner text faster and make your work sound more trustworthy.
https://github.com/hardikpandya/stop-slop
This skill helps you remove AI-like writing patterns so your prose sounds more natural, direct, and human. It gives clear rules for cutting banned phrases, avoiding cliché structures, using active voice, and checking quality, which helps you write cleaner text faster and make your work sound more trustworthy.
https://github.com/hardikpandya/stop-slop
GitHub
GitHub - hardikpandya/stop-slop: A skill file for removing AI tells from prose
A skill file for removing AI tells from prose. Contribute to hardikpandya/stop-slop development by creating an account on GitHub.
#shell #agent #ai #claude #claude_code #codex #coding #design #frontend #lowcode #nocode #skill #skills #vibecoding
Taste Skill is a set of portable AI design tools that helps agents build better-looking interfaces with stronger layout, spacing, typography, and motion instead of plain, boilerplate UIs. It also includes image tools for web, mobile, and brand reference boards, so you can create polished design frames, then turn them into code with tools like Codex, Cursor, or Claude Code. The benefit is faster work with better visual quality and less repetitive design output.
https://github.com/Leonxlnx/taste-skill
Taste Skill is a set of portable AI design tools that helps agents build better-looking interfaces with stronger layout, spacing, typography, and motion instead of plain, boilerplate UIs. It also includes image tools for web, mobile, and brand reference boards, so you can create polished design frames, then turn them into code with tools like Codex, Cursor, or Claude Code. The benefit is faster work with better visual quality and less repetitive design output.
https://github.com/Leonxlnx/taste-skill
GitHub
GitHub - Leonxlnx/taste-skill: Taste-Skill - gives your AI good taste. stops the AI from generating boring, generic slop
Taste-Skill - gives your AI good taste. stops the AI from generating boring, generic slop - Leonxlnx/taste-skill