MiMo-V2.5-TP3-NVFP4-KV-3xDGX-Spark. MiMo-V2.5 Omni · TP=3 on 3x DGX Spark · NVFP4 4-bit KV cache → ~3.5x more 1M-context KV pool (3.06M→10.59M tokens) vs FP8

github.com/tonyd2wild/MiMo-V2.5-TP3-NVFP4-KV-3xDGX-Spark

Vaya's read on this project

Problem, audience, market, and the verdict — sign in to see it.

Updates

No recent activity.