R_REDDYX.XYZ
AIC★ 2,559⑂ 194 forksApache-2.0

sqliteai/warp

Warp streaming output

Warp enables inference of the Kimi K3 model with 2.78 trillion parameters, streaming activated weights directly from NVMe, bypassing RAM limitations. The engine is written in pure C without external dependencies and easily embeds into projects.

// KEY FEATURES

  • Streaming loading of weight layers from NVMe, allowing models larger than RAM to run.
  • Library-independent pure C code, easily embeddable into any projects.
  • Support for the full Kimi K3 model (2.78 trillion parameters) for real-time inference.
  • Minimal latency thanks to direct streaming weight transfer without intermediate buffering.
Open on GitHub →

New repositories every 30 minutes

REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.

Join on Telegram
← Full catalog·Full index

// SIMILAR REPOSITORIES

Accelerate Deep Analysis★ 2559OpenUsage — Subscription Management★ 2560DATA TRANSMISSION BRIDGE★ 2562Claude Code for Terminal★ 2554DOCMD — DOCUMENTATION IN SECONDS★ 2565NITROSTACK: FULL AI STACK★ 2565

← FULL CATALOG

sqliteai/warp — Warp streaming output | REDDYX AI