⚡ Each donation = another big MoE quantized

I host 30+ free APEX MoE quantizations as independent research. My only local hardware is an NVIDIA DGX Spark (122 GB unified memory), enough for ~30-50B-class MoEs, but bigger ones (200B+) require rented compute on H100/H200/Blackwell, typically $20-100 per quant.
If APEX quants are useful to you, your support directly funds those bigger runs.

🎉 Patreon (Monthly)  |  ☕ Buy Me a Coffee  |  ⭐ GitHub Sponsors

Laguna-XS-2.1 — APEX GGUF

APEX (Adaptive Precision for EXpert Models) quantizations of poolside/Laguna-XS-2.1 — poolside's Laguna XS.2 Mixture-of-Experts model for coding and agentic software engineering.

Brought to you by the LocalAI team | APEX Project | Technical Report

Requires a recent llama.cpp with Laguna support (PR #25165). Older builds cannot load arch=laguna.

Available Files

What is APEX?

APEX is a quantization strategy for Mixture-of-Experts (MoE) models. It classifies tensors by role (routed expert, shared expert, attention, dense FFN) and applies a layer-wise precision gradient — edge layers get higher precision, middle layers compress more aggressively. I-variants use diverse imatrix calibration (chat, code, reasoning, tool-calling, agentic traces, Wikipedia).

In MoE models the routed-expert FFN tensors dominate the weight budget but only ~8/256 experts fire per token, so APEX compresses middle-layer routed experts hardest while preserving edge layers, attention, and the always-active shared expert.

APEX layout for Laguna

Laguna XS.2 has a structure APEX handles explicitly:

• Layer 0 is a leading dense FFN (no experts) — pinned to Q8_0, since every token traverses it.
• Layers 1–39 are MoE — 256 routed experts + a shared expert, 8 active per token, sigmoid gating.
• Shared expert (ffn_*_shexp) kept at Q8_0 on every tier (always active).
• Routed experts follow the 5+5 symmetric edge gradient (higher precision at the first/last layers, most aggressive in the middle).
• Router (ffn_gate_inp), norms and the exp_probs_b gating bias stay at full precision.

Architecture

• Model: Laguna-XS-2.1 (LagunaForCausalLM, arch laguna)
• Layers: 40 (1 dense + 39 MoE) · Experts: 256 routed + 1 shared (8 active)
• Attention: 48 heads / 8 KV, per-layer output gate, hybrid full + sliding-window, YaRN rope
• Vocab: 100352 · text-only
• Calibration: v1.3 diverse dataset

Run with LocalAI

local-ai run mudler/Laguna-XS-2.1-APEX-GGUF@Laguna-XS-2.1-APEX-I-Balanced.gguf

Credits

APEX is brought to you by the LocalAI team. Built on llama.cpp. Base model by poolside.