File size: 5,996 Bytes

559d731

---
license: mit
base_model: litert-community/NAFNet-GoPro-width32-LiteRT
base_model_relation: quantized
library_name: openvino
pipeline_tag: image-to-image
tags:
- openvino
- onnx
- intel-npu
---

# NAFNet-GoPro-width32-OpenVINO

NAFNet-GoPro-width32 for Intel NPUs, GPUs and CPUs: nafnet_fp16 from [litert-community/NAFNet-GoPro-width32-LiteRT](https://huggingface.co/litert-community/NAFNet-GoPro-width32-LiteRT/tree/84fa01bda5ea09a562e83fa1863b951453b5ccf0) as ONNX (tf2onnx 1.17.0, opset 17), OpenVINO IR with fp32 weights, and a precompiled Intel NPU blob. On the NPU of a Core Ultra 9 386H the blob runs in 25.21 ms per run, numpy input to numpy output.

Scripts, raw records, and the same table for all 52 graphs of LiteRT-Models: https://github.com/john-rocky/windows-zoo

## Files

| file | bytes | sha256 |
|---|---|---|
| `nafnet_fp16.onnx` | 75,757,684 | `9e3e1699ffddf39a17a9572ef7da7030da6b96827f681b60e59c041a5523d6e3` |
| `openvino/nafnet_fp16.xml` | 4,296,286 | `ede5ca4627bd5631dc270bbb29f13f358b141ea9427cb21e3ce95b64ecaebfe6` |
| `openvino/nafnet_fp16.bin` | 74,309,056 | `8516d9ba48287408a63502e9070f26efeb79c3f764f42443a99d965cc031fd95` |
| `npu/nafnet_fp16.blob` | 49,107,073 | `5365c926cd8083edc582e26536c1584d788cab5823cd4cb86394a73b21574430` |

The NPU blob was compiled with OpenVINO 2026.4.1, Intel NPU driver 32.0.100.4841 (OpenVINO reports NPU_DRIVER_VERSION 1004841, NPU_COMPILER_VERSION 524292). If `import_model` refuses a blob on another driver or OpenVINO version, compile the IR instead (the commented line in the snippet).

## How it was measured

MSI Stealth 16 AI+ (Core Ultra 9 386H, RTX 5080 Laptop GPU at a 95 W power limit, Windows 11 build 26200.9711, on AC power), 2026-11-06 to 2026-11-07. On the laptop, each row is one process: load, one run, 5 warmup runs, then the median of 30 timed runs (2 warmup + 10 timed runs when the first run takes more than 3 s), numpy input to numpy output, every runtime at its default settings, one process at a time, each started after the CPU, the RTX and the Intel GPU had gone idle (the thresholds are in METHOD.md of the GitHub repo). Output check: each output against the CPU output in fp32 on the same input (ONNX Runtime's CPU EP); PASS = cosine ≥ 0.99, no NaN or Inf, and top-1 agreement ≥ 0.99 for class scores. The Snapdragon rows are Qualcomm AI Hub profile jobs of the same ONNX files (their own timer).

## Speed and output check

| processor | runtime | file | ms | load s | output check |
|---|---|---|---|---|---|
| Intel AI Boost NPU | OpenVINO 2026.4.1 | `npu/` blob (`import_model`) | 25.21 | 0.04 | PASS (cos ≥ 0.9999) |
| Intel AI Boost NPU | ONNX Runtime 1.31.1 + OpenVINO EP 1.8.1 | `.onnx`, compiled when the session loads | 26.90 | 59.11 | PASS (cos ≥ 0.9999) |
| Intel Graphics (iGPU) | ONNX Runtime 1.31.1 + OpenVINO EP 1.8.1 | `.onnx` | 29.01 | 8.25 | PASS (cos ≥ 0.9999) |
| RTX 5080 Laptop (95 W limit) | ONNX Runtime 1.31.1, CUDA EP | fp16 copy of the `.onnx` (not in this repo) | 49.84 | 2.30 | PASS (cos ≥ 0.9999) |
| RTX 5080 Laptop (95 W limit) | ONNX Runtime 1.31.1, CUDA EP | `.onnx` | 26.87 | 1.48 | PASS (cos ≥ 0.9999) |
| Core Ultra 9 386H CPU | ONNX Runtime 1.31.1, CPU EP | `.onnx` | 639.23 | 0.52 | reference |
| Snapdragon X2 Elite NPU | Qualcomm AI Hub, precompiled QNN ONNX | `.onnx` compiled by AI Hub | 19.06 | 0.41 | not compared (AI Hub profile job, 2026-10-30) |
| Snapdragon X2 Elite NPU | Qualcomm AI Hub, ONNX Runtime QNN EP | `.onnx`, compiled when it loads | 36.05 (4 layers on the CPU) | 47.94 | not compared (AI Hub profile job, 2026-10-30) |
| Galaxy S26 NPU | Qualcomm AI Hub, precompiled QNN ONNX | `.onnx` compiled by AI Hub | 18.33 | 0.23 | not compared (AI Hub profile job, 2026-10-30) |

The ONNX file against the source `.tflite` (LiteRT interpreter on the CPU, the same input): PASS, cosine ≥ 0.9999 for every output.

## Run it on the Intel NPU

Download this repo, then run the snippets from its folder:

```sh
pip install huggingface_hub
hf download mlboydaisuke/NAFNet-GoPro-width32-OpenVINO --local-dir NAFNet-GoPro-width32-OpenVINO
cd NAFNet-GoPro-width32-OpenVINO
```

OpenVINO (`pip install openvino==2026.4.1`):

```python
import numpy as np
import openvino as ov

core = ov.Core()
if "NPU" not in core.available_devices:
    raise SystemExit(f"no Intel NPU visible to OpenVINO: {core.available_devices}")
with open("npu/nafnet_fp16.blob", "rb") as f:
    compiled = core.import_model(f.read(), "NPU")  # precompiled: no compile at load
# or compile the IR on this machine: compiled = core.compile_model("openvino/nafnet_fp16.xml", "NPU")
# replace the zeros with your preprocessed input (float32, the shapes printed by compiled.inputs)
feeds = {i.get_any_name(): np.zeros(list(i.shape), np.float32) for i in compiled.inputs}
outputs = compiled(feeds)
print({o.get_any_name(): outputs[o].shape for o in compiled.outputs})
```

ONNX Runtime with the OpenVINO EP (`pip install onnxruntime==1.31.1 onnxruntime-ep-openvino==1.8.1`):

```python
import numpy as np
import onnxruntime as ort
import onnxruntime_ep_openvino as ov_ep

ort.register_execution_provider_library(ov_ep.get_ep_name(), ov_ep.get_library_path())
npu = [d for d in ort.get_ep_devices()
       if d.ep_name == ov_ep.get_ep_name() and d.ep_metadata.get("ov_device") == "NPU"]
if not npu:
    raise SystemExit("no Intel NPU visible to the OpenVINO EP")
so = ort.SessionOptions()
so.add_provider_for_devices(npu, {})
sess = ort.InferenceSession("nafnet_fp16.onnx", sess_options=so)  # compiles for the NPU here
# replace the zeros with your preprocessed input (float32)
feeds = {i.name: np.zeros(i.shape, np.float32) for i in sess.get_inputs()}
outputs = sess.run(None, feeds)
print([o.shape for o in outputs])
```

## Licence

MIT, the licence of [litert-community/NAFNet-GoPro-width32-LiteRT](https://huggingface.co/litert-community/NAFNet-GoPro-width32-LiteRT). The source card names the upstream model as `megvii-research/NAFNet` (not a Hugging Face repo).