MiniCPM5-2B-OpenNPU2-NPU2

OpenFlowLM Q4NX conversion of openbmb/MiniCPM5-2B for AMD XDNA NPU inference.

This repository contains a quantized Q4NX port of the model, compiled for the OpenFlowLM (OFLM) runtime. It is not a GGUF file.

Item	Value
Source model	`openbmb/MiniCPM5-2B`
Source GGUF	`MiniCPM5-2B.i1-Q4_1.gguf`
Weights	`model.q4nx` (1.81 GB)
Modality	language
OFLM version	`0.1.0`
Converted	2026-09-14

Install and run

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

pip install oflm-add or uv tool install oflm-add

uv tool install oflm-add
oflm-add Atomic-Germ/MiniCPM5-2B-OpenNPU2-NPU2 --family qwen3.5 --xclbin-from MiniCPM5-2B-OpenNPU2-NPU2
OFLM_CONFIG_PATH="$HOME/.config/oflm/model_list.json" OFLM_XCLBIN_PATH="$HOME/.config/oflm" oflm run MiniCPM5-2B-OpenNPU2-NPU2

Files

File	Description
`model.q4nx`	Quantized weights (Q8_0 / Q4_1 / BF16)
`config.json`	OFLM runtime configuration
`tokenizer.json`	Tokenizer vocabulary
`tokenizer_config.json`	Tokenizer configuration
`chat_template.jinja`	Chat template

Source model card

See the original model card: openbmb/MiniCPM5-2B

加载中…

LD_LIBRARY_PATH=/home/beta1/qwen38-release/tools/llama-b10442-cuda /home/beta1/qwen38-release/tools/llama-b10442-cuda/llama-bench -m /mnt/pf/pf-deep2/MiniCPM5-2B-OpenNPU2/MiniCPM5-2B.i1-Q4_K_M.gguf -p 512 -n 128 -ngl 99 -fa 1