---
license: apache-2.0
base_model: mlboydaisuke/Depth-Anything-3-Small-LiteRT
base_model_relation: quantized
library_name: openvino
pipeline_tag: depth-estimation
tags:
- openvino
- onnx
- intel-npu
---

# Depth-Anything-3-Small-OpenVINO

Depth-Anything-3-Small for Intel NPUs, GPUs and CPUs: da3_small_gpu_fp16 from [mlboydaisuke/Depth-Anything-3-Small-LiteRT](https://huggingface.co/mlboydaisuke/Depth-Anything-3-Small-LiteRT/tree/15344dc67dca2065dbb5fa625c5e8e6e5a228b57) 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 12 FULLY_CONNECTED layers of da3_small_gpu_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 97.05 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 |
|---|---|---|
| `da3_small_gpu_fp16.onnx` | 104,425,500 | `8e678f8b980878ffaac506070e9c6995b359c626a9ed31ae00c9de0ed0902a8e` |
| `openvino/da3_small_gpu_fp16.xml` | 2,173,241 | `904c6fca9a8e15f4f7dae77e0de669f68221dd11d73fadfbc909d3a39aaf8aa8` |
| `openvino/da3_small_gpu_fp16.bin` | 103,246,984 | `6f7428dfc08a879a1e3e3bbf00dd76b25087512a141f9d1046afd9db55f75572` |
| `npu/da3_small_gpu_fp16.blob` | 111,669,377 | `fd72ff655c6e5e22eebc32e1039bec273e68d0edbb83e12130d0977f4ecd4aeb` |

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`) | 97.05 | 0.08 | PASS (cos ≥ 0.9999) |
| Intel AI Boost NPU | ONNX Runtime 1.31.1 + OpenVINO EP 1.8.1 | `.onnx`, compiled when the session loads | 96.67 | 186.59 | PASS (cos ≥ 0.9999) |
| Intel Graphics (iGPU) | ONNX Runtime 1.31.1 + OpenVINO EP 1.8.1 | `.onnx` | 158.36 | 4.76 | 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) | 25.93 | 1.48 | PASS (cos ≥ 0.9999) |
| RTX 5080 Laptop (95 W limit) | ONNX Runtime 1.31.1, CUDA EP | `.onnx` | 43.41 | 1.35 | PASS (cos ≥ 0.9999) |
| Core Ultra 9 386H CPU | ONNX Runtime 1.31.1, CPU EP | `.onnx` | 1345.45 | 0.46 | 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:

```sh
pip install huggingface_hub
hf download mlboydaisuke/Depth-Anything-3-Small-OpenVINO --local-dir Depth-Anything-3-Small-OpenVINO
cd Depth-Anything-3-Small-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/da3_small_gpu_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/da3_small_gpu_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("da3_small_gpu_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

Apache-2.0, the licence of [mlboydaisuke/Depth-Anything-3-Small-LiteRT](https://huggingface.co/mlboydaisuke/Depth-Anything-3-Small-LiteRT). The upstream model: [depth-anything/DA3-SMALL](https://huggingface.co/depth-anything/DA3-SMALL).