NuExtract3-4B-NPU2

IF YOU USE COMMUNITY QWEN MODELS DO NOT UPGRADE TO FLM v1.0.2+

NuExtract3-4B - Q4NX for FastFlowLM (AMD Ryzen AI XDNA2)

A text-extraction Qwen3.5-4B fine-tune, converted to Q4NX for FastFlowLM.

What is Q4NX?

Q4NX is FastFlowLM's native packed-quantization format - a rearranged Q4_1 layout tuned for the NPU matrix engine's tile sizes and memory access patterns. It is not a GGUF file and it does not run on llama.cpp or Ollama; it is meant exclusively for the FastFlowLM engine on AMD Ryzen AI NPUs.

Requirements

• FastFlowLM >= 0.9.45 (flm CLI)
• AMD Ryzen AI processor with XDNA2 (NPU2) - Strix Point / Ryzen AI 300 series or later
• Linux with the XRT NPU stack installed
• ~16 GB of unified system memory (Q4NX weights + activations + KV cache)

Files

Install and run

This repository works with flm-add, a small installer that copies the model into the FastFlowLM user directory and registers the tag. It never modifies the system FastFlowLM install.

pip install flm-add or uv tool install flm-add

uv tool install flm-add
flm-add Atomic-Germ/NuExtract3-4B-NPU2 --tag nuextract3:4b --family qwen3.5
FLM_CONFIG_PATH="$HOME/.config/flm/model_list.json" FLM_XCLBIN_PATH="$HOME/.config/flm" flm run nuextract3:4b

Kernels

FastFlowLM's NPU kernels (xclbins) are closed source and are not shipped in this repository. flm-add.py links the kernels of the official qwen3.5:4b model (Qwen3.5-4B-NPU2), because this model shares the same engine family (qwen3.5) and architecture.

Model

• Registry tag: nuextract3:4b
• Engine family: qwen3.5
• Kernel source: Qwen3-4B-NPU2
• Context length: 32768 tokens (from config)
• model.q4nx size: 4.3 GB
• Base model: Qwen/Qwen3.5-4B
• License: apache-2.0

---

Original model card

See the upstream model card for training details, benchmarks, and upstream usage. This repository only contains the Q4NX conversion for FastFlowLM.

• Upstream card: numind/NuExtract3