These hardware-specific .rkllm files require a compatible Rockchip RKLLM runtime. They are not Transformers checkpoints and cannot be loaded directly with Transformers, llama.cpp, or Ollama.

Base model

Upstream model: Qwen/Qwen2.5-1.5B-Instruct
License: Apache-2.0
Model type: LLM (text only)

Conversion and variants

Toolkit version

RKLLM Toolkit: v1.2.3

Use a file built for the exact target SoC.

Target	Quantization	File	SHA256
RK3576	W4A16	Qwen2.5-1.5B-Instruct_RK3576_w4a16.rkllm	`70b4b3c54221892cc403aa877612e0bfdc24abeda5e0011ee3a2e9a3c24e332c`
RK3576	W8A8	Qwen2.5-1.5B-Instruct_RK3576_w8a8.rkllm	`a2ab655ca7e6a3d3626a9a93e69a3a95660cedd7d4820ddf041371b45f4a6c94`
RK3588	W8A8	Qwen2.5-1.5B-Instruct_RK3588_w8a8.rkllm	`e296d41ff1ff64f86e293227f15e35520dbeacebb74d21592a0790919691ca2d`

The repository also includes Qwen2.5-1.5B-Instruct_data_quant.json, used as calibration data during conversion.

Usage

bash

hf download HanzoHuang/Qwen2.5-1.5B-Instruct-RKLLM \
  RK3576/Qwen2.5-1.5B-Instruct_RK3576_w4a16.rkllm \
  --local-dir Qwen2.5-1.5B-Instruct-RKLLM

Use the Qwen2.5 Instruct chat template with the RKLLM runtime. For Docker deployment, see Hanzo-Huang/rkllm-docker.

Limitations

These are target-specific converted artifacts. Validate quality and runtime compatibility on your Rockchip device.

Acknowledgements

Thanks to the Qwen Team, Rockchip, and the RKLLM community.