Description

A high-performance inference engine forked from Project Maya, optimized for running the 428B parameter MiniMax-M3.1-Flash model locally. It utilizes Unsloth 2-bit GGUF weights, a DSpark drafter, and a custom expert cache that offloads weights to system RAM to achieve 5x prefill and 4.6x decode speedups on consumer hardware.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.