Metric3D-v2-OpenVINO

Metric3D-v2 for Intel NPUs, GPUs and CPUs: metric3d_fp16 from mlboydaisuke/Metric3D-v2-LiteRT as ONNX (tf2onnx 1.17.0, opset 17), OpenVINO IR with fp32 weights, and a precompiled Intel NPU blob. tf2onnx 1.17.0 drops the bias of 24 FULLY_CONNECTED layers of metric3d_fp16; these files were converted with a one-line fix of that handler (convert/patch_tf2onnx.py in the GitHub repo), and their outputs match the .tflite. On the NPU of a Core Ultra 9 386H the blob runs in 80.40 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
metric3d_fp16.onnx	153,956,920	05beedc220e804dd418a4e75faa20d001505ef64ca79ae558c30290571b2e185
openvino/metric3d_fp16.xml	5,556,781	4df486a82b7b0ab2ee49dd1c09e6155b24db3bd48e22ba2407f726edadd3d006
openvino/metric3d_fp16.bin	150,206,156	f8c415fbcc763f3772229e74865e2641151f4d21224a096087bf457129684039
npu/metric3d_fp16.blob	105,549,953	fef43f864791c18a60e802d3d2d3abb12c373bd0ad248d19ca25f898f83d340d

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), 2026-11-07. 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. No Snapdragon rows: Qualcomm AI Hub measured the file before the bias fix (its numbers are in the GitHub table).

Speed and output check

processor	runtime	file	ms	load s	output check
Intel AI Boost NPU	OpenVINO 2026.4.1	npu/ blob (import_model)	80.40	0.08	PASS (cos ≥ 0.9991)
Intel AI Boost NPU	ONNX Runtime 1.31.1 + OpenVINO EP 1.8.1	.onnx, compiled when the session loads	69.92	314.25	PASS (cos ≥ 0.9991)
Intel Graphics (iGPU)	ONNX Runtime 1.31.1 + OpenVINO EP 1.8.1	.onnx	56.25	10.06	PASS (cos ≥ 0.9988)
RTX 5080 Laptop (95 W limit)	ONNX Runtime 1.31.1, CUDA EP	fp16 copy of the .onnx (not in this repo)	58.00	1.66	PASS (cos ≥ 0.9991)
RTX 5080 Laptop (95 W limit)	ONNX Runtime 1.31.1, CUDA EP	.onnx	57.03	1.67	PASS (cos ≥ 0.9999)
Core Ultra 9 386H CPU	ONNX Runtime 1.31.1, CPU EP	.onnx	1367.07	0.76	reference

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:

pip install huggingface_hub
hf download mlboydaisuke/Metric3D-v2-OpenVINO --local-dir Metric3D-v2-OpenVINO
cd Metric3D-v2-OpenVINO

OpenVINO (pip install openvino==2026.4.1):

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/metric3d_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/metric3d_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):

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("metric3d_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

BSD-2-Clause, the licence of mlboydaisuke/Metric3D-v2-LiteRT. The source card names the upstream model as yvanyin/metric3d (not a Hugging Face repo).