microsoft / microsoft/onnxruntime

[OVEP][NPU]compile model will hang all the time and never return when running model florence-2-base-vision-encoder-fp16 with OpenVINO-NPU

Open
#26,643 3 comments 0 reactions 1 assignee View on GitHub

Nobody has claimed this yet.

ep:OpenVINO
Dominant language
C++
Stars
21.9k
Forks
4.2k
Avg merge
4d 11h
Merged PRs (30d)
184

Description

### Describe the issue

When running florence-2-base-vision-encoder-fp16 model with onnxruntime_perf_test tool OpenVINO-NPU, it will hang in core.compile_model in https://github.com/microsoft/onnxruntime/blob/main/onnxruntime/core/providers/openvino/ov_interface.cc#L165 and never return, which blocks the program to run

### To reproduce

1. Build onnxruntime code with "--use_openvino" flag
2. Download florence-2-base-vision-encoder-fp16 from https://huggingface.co/onnx-community/Florence-2-base/blob/main/onnx/vision_encoder_fp16.onnx
3. Run command "onnxruntime_perf_test.exe -I -r 50 -e openvino -i "device_type|NPU" -f batch_size:1 -f height:768 -f width:768 vision_encoder_fp16.onnx"

### Urgency

_No response_

### Platform

Windows

### OS Version

Win11

### ONNX Runtime Installation

Built from Source

### ONNX Runtime Version or Commit ID

977efe4788b2ee24371523b5fa14dd02efcd4942

### ONNX Runtime API

C++

### Architecture

X64

### Execution Provider

OpenVINO

### Execution Provider Library Version

OpenVINO 2025.3

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.