#cplusplus
FlashMLA is DeepSeek's optimized attention library that makes AI models run faster and use less memory. It works with advanced NVIDIA GPUs to speed up how language models process information, achieving up to 660 trillion floating-point operations per second. The library supports both dense and sparse attention modes, meaning it can focus on important tokens while skipping less relevant ones, reducing computational waste. For you, this means faster AI responses, lower costs for running large language models, and better performance on tasks like chatbots and code generation. The technology is open-source and integrates with popular AI frameworks like PyTorch and Hugging Face, making it accessible for developers building next-generation AI applications.
https://github.com/deepseek-ai/FlashMLA
kitty
The fast, feature-rich, GPU based #terminal emulator
Uses GPU and SIMD vector CPU instructions for best in class
Uses threaded rendering for absolutely minimal latency
Performance tradeoffs can be tuned
Capable Scriptable Composable Cross-platform Innovative
To get started see Quickstart.
https://sw.kovidgoyal.net/kitty/
With software, such as "Kitty", it is possible to read documents, images and other graphical formats with these software:
https://github.com/dsanson/termpdf.py
http://www.kraxel.org/blog/linux/fbida/
https://github.com/itsjunetime/tdf
It means, that we no longer need GTK+ (which dropped support for X11), Qt and other graphical toolkits in order to have a full operational computer.
qrcp, transfer files over wifi from your computer to your mobile device by scanning a QR code without leaving the terminal.
#tools#terminal
@thedevs
https://kutt.it/fycGgm
eDEX-UI, a science fiction terminal emulator designed for large touchscreens that runs on all major OSs.
#tools#terminal
@thedevs
https://kutt.it/IaR1au