
How TatuEngine stopped blowing its memory budget with ternary weights
A technical post-mortem of how we replaced TatuEngine's weight format with ternary blocks and got the model back inside consumer GPU memory, without losing inference coherence.
4 posts

A technical post-mortem of how we replaced TatuEngine's weight format with ternary blocks and got the model back inside consumer GPU memory, without losing inference coherence.

BitMamba-2 1B's in-engine inference representation was unworkable in memory. A compact ternary block codec delivered orders of magnitude of savings, with the lm_head diverging mid-validation.

Born from the absurd question 'what if we used RT Cores to process neural networks?', killed the dream on WSL2, and became a hybrid CPU/GPU engine with BitMamba-2 1B, radical compression, an agent with 8 subsystems and a security sandbox — 302 commits later.

The journey (and the craters) of training a Mamba-2 with 1.4B parameters from scratch on an RTX 3060 — warmup, SFT, catastrophic eval, and Phase 4 of the resurrection.