Cross-platform desktop application development framework Electron

electron/electron

Stars: 109.3k License: MIT

Electron is an open source framework based on Node.js and Chromium that allows cross-platform desktop applications to be written using JavaScript, HTML and CSS. It is used by many applications such as the Atom editor. The project has the following core advantages:

Cross-platform: Electron provides binaries for the three major operating systems: macOS, Windows and Linux.
Simplify the development process: The Electron Fiddle tool makes it easy to build, run, and package small experimental Electron code, and try different versions of Electron.
Rich documentation resources: The official website provides all documentation about Electron.

d2l-ai/d2l-zh

Stars: 46.5k License: Apache-2.0

This open source project represents our attempt to teach readers the concepts, background, and code; we will illustrate the critical thinking required to analyze the problem, the mathematics required to solve the problem, and the implementation of the solution all in one place. Engineering skills required for the program.
Our goal is to create a unified resource that achieves the following goals:

Free and available to everyone online;
Provide enough technical depth to help readers actually become deep learning application scientists: not only understand the mathematical principles, but also be able to implement and continuously improve methods;
Contains working code that shows readers how to solve problems in practice. This not only directly corresponds to mathematical formulas into actual code, but also allows you to modify the code, observe the results, and gain experience in time;
Allows us and the entire community to continuously iterate content quickly to keep up with the still rapidly evolving field of deep learning;
This is supplemented by a forum containing questions and answers on technical details, allowing everyone to answer each other's questions and exchange experiences.

firezone/firezone

Stars: 5.0k License: Apache-2.0

Firezone is a self-hosted VPN server and Linux firewall project.

Manage remote access through an intuitive web interface and CLI utility.
Deploy on your own infrastructure to control network traffic.
Built on WireGuard®, it is stable, performant and lightweight.

core advantages:

Fast: 3 to 4 times faster than OpenVPN using WireGuard®
SSO integration: authenticate using any identity provider with OpenID Connect (OIDC) connector
Containerization: all dependencies are packaged in Docker
Easy to use: It takes just minutes to set up and can be managed via a simple command line interface.

vllm-project/vllm

Stars: 4.4k License: Apache-2.0

vLLM is a fast and easy-to-use library for LLM inference and serving.

The following are the core benefits of vLLM:

Featuring state-of-the-art inference throughput
Use PagedAttention to efficiently manage attention key value memory
Perform continuous batch processing to process incoming requests
Optimized CUDA kernel

vLLM has the following key features:

Seamlessly integrated with the popular HuggingFace model, flexible and easy to use.
High-throughput services with various decoding algorithms, including parallel sampling, beam search, and more.
Tensor parallelism support for distributed inference.
vLMM is fully compatible with the OpenAI API Server specification.

In addition, vLLM can seamlessly support many Huggingface models. For example, BLOOM, GPT2, GPT BigCode, etc., and surpass HuggingFace Transformers (HF) and Text Generation Inference (TGI) in terms of performance, with speeds up to 24 times that of HF and 3.5 times that of TGI.

continuedev/continue

Stars: 3.4k License: Apache-2.0

Continue is an extension plug-in for VS Code that introduces the powerful functions of ChatGPT into the IDE. The main functions of the project include:

Answer coding questions: You can highlight code snippets and ask Continue to get another perspective.
Edit code in natural language: You can highlight code segments and instruct Continue to refactor.
Generate files from scratch: Open a blank file and let Continue create new Python scripts, React components, and more.
GPT-4Using the OpenAI API, the and GPT3.5-turbomodels are used by default , but can also be configured to use different LLMs (such as local or private models).

RayVentura/ShortGPT

Stars: 3.3k License: NOASSERTION

ShortGPT is an experimental AI framework for automating short video content creation.

The main function:

Automated editing framework: Simplifies the video creation process using a video editing language for LLM.
Scripts and Tips: A variety of pre-prepared scripts and tips are provided to automate multiple LLM editing processes.
Voice dubbing/content creation: Supports multiple languages including English, Spanish, Arabic, etc., and more than 30 other languages are available (via EdgeTTS).
Subtitle generation: Automatically generate video subtitles.
Material acquisition: Obtain images and video materials from the Internet and connect to the network or Pexels API to operate as needed.
Memory and Persistence: Use TinyDB to ensure long-term preservation of autoedited variables.