

mitmproxy/mitmproxyhttps://github.com/mitmproxy/mitmproxy
Stars: <span>38.2k</span> License: <span>MIT</span>

mitmproxy is an interactive, SSL/TLS-capable HTTP proxy designed for penetration testers and software developers.
- Supports HTTP/1, HTTP/2, and WebSockets protocols
- Provides a command-line version mitmdump, similar to tcpdump for HTTP traffic analysis
- Features a web-based interface mitmweb for user-friendly operation and monitoring
- Offers extensive documentation and tutorials to help users get started
mailpile/Mailpilehttps://github.com/mailpile/Mailpile
Stars: <span>8.8k</span> License: <span>NOASSERTION</span>
Mailpile is a modern, fast email client with user-friendly encryption and privacy features.
- Provides a clear and elegant user interface to enhance user experience
- Supports PGP encryption and signing for sending and receiving emails
- Based on web technologies, can be used as a local desktop application or a remote webmail
- Includes a fast search engine suitable for handling large volumes of emails
- Allows organizing emails by tags and can automatically label incoming messages
FasterDecoding/Medusahttps://github.com/FasterDecoding/Medusa
Stars: <span>2.5k</span> License: <span>Apache-2.0</span>

Medusa is a simple framework designed to accelerate LLM generation with multiple decoding heads.
- Provides multi-head decoding technology to enhance generation speed.
- Efficient training process suitable for users with limited resources.
- Maintains the original model unchanged, only fine-tuning the added decoding heads.
- Supports self-distillation, allowing Medusa to be applied to any fine-tuned LLM without the need for original training data.
- Achieves approximately 2x speedup at a batch size of 1.
evolutionary-architecture/evolutionary-architecture-by-examplehttps://github.com/evolutionary-architecture/evolutionary-architecture-by-example
Stars: <span>2.3k</span> License: <span>MIT</span>

evolutionary-architecture-by-example is a project that guides users through understanding .NET software architecture, exploring the interactions between modular monoliths, microservices, domain-driven design, and various architectural patterns through storytelling.
- Provides a clear and understandable evolution process of software architecture
- Combines various architectural approaches to adapt to different needs
- Divided into four chapters, covering solutions from simple to complex
- Emphasizes the importance of maintainability and growth potential
gnobitab/InstaFlowhttps://github.com/gnobitab/InstaFlow
Stars: <span>1.3k</span> License: <span>MIT</span>

InstaFlow is an ultra-fast one-step image generator that utilizes flow correction technology to achieve image quality close to stable diffusion while significantly reducing computational resource requirements.
- Ultra-fast inference: generates in one step, with an inference time of about 0.1 seconds, saving approximately 90% of the time compared to the original stable diffusion.
- High-quality output: generates detail-rich images with an FID comparable to state-of-the-art text-to-image GANs.
- Simple and efficient training: requires only supervised training and utilizes pre-trained models, taking only 199 A100 GPU days to complete.
