Posts by Bilawal Riaz

• MiniMax-01: 4 Million Token Context Window!

  Lightning Attention: 4M token window, 456B params. Open-source LLM matches GPT-4, with affordable API. Redefines long-context processing.

  B Bilawal Riaz

  January 17, 2025

  4 min read

  llm

  ai

  programming

  machine-learning

  open-source

• Titans: Revolutionizing Long-Context Processing in LLMs

  Innovative architecture combining short-term attention and long-term memory, handling 2M+ tokens. Outperforms GPT-4 in long-context tasks.

  B Bilawal Riaz

  January 17, 2025

  4 min read

  llm

  ai

  long-context

  machine-learning

  neural-architecture

• Kokoro: A Tiny TTS Model That Beats the Giants

  82M parameters • Beats 1B+ models • Apache 2.0 licensed • Self-hostable • Multiple voices • Easy deployment

  B Bilawal Riaz

  January 15, 2025

  3 min read

  machine-learning

  text-to-speech

  open-source

  fastapi

  docker

  tts

• Sup world

  A tour of this website, its tech stack, and what you can expect to find here.

  B Bilawal Riaz

  January 14, 2025

  2 min read

  astro

  react

  tailwind