---
title: YOLOE-Segmentation
canonical_url: "https://www.modelscope.cn/models/Qualcomm-AI-Hub/YOLOE-Segmentation"
md_url: "https://www.modelscope.cn/models/Qualcomm-AI-Hub/YOLOE-Segmentation.md"
repository: Qualcomm-AI-Hub/YOLOE-Segmentation
last_updated: 2026-10-16
license: agpl-3.0
pipeline_tag: image-segmentation
tasks:
  - image-segmentation
downloads: 0
stars: 0
tags:
  - real_time
  - android
---

# YOLOE-Segmentation

> YOLOE-Segmentation - Qualcomm-AI-Hub 在 ModelScope 开源的模型。YOLOE-Segmentation: Optimized for Qualcomm Devices

Qualcomm-AI-Hub/YOLOE-Segmentation 是 ModelScope 魔搭社区上的image-segmentation模型，采用 agpl-3.0 许可。

- **Repository**: Qualcomm-AI-Hub/YOLOE-Segmentation
- **License**: agpl-3.0
- **Tasks**: image-segmentation
- **Tags**: real_time, android
- **Downloads**: 0
- **Stars**: 0
- **Last updated**: 2026-10-16

Source: https://www.modelscope.cn/models/Qualcomm-AI-Hub/YOLOE-Segmentation

---

![](https://china-qaihub-public-assets.s3.cn-northwest-1.amazonaws.com.cn/qai-hub-models/models/yoloe_seg/web-assets/model_demo.png)

# YOLOE-Segmentation: Optimized for Qualcomm Devices

Ultralytics YOLOE is a machine learning model that predicts bounding boxes, segmentation masks and classes of objects in an image.

This is based on the implementation of YOLOE-Segmentation found [here](https://github.com/THU-MIG/yoloe).
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models-cn/blob/main/src/qai_hub_models/models/yoloe_seg) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).

Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.cn) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) to run these models on a hosted Qualcomm® device.

## Getting Started
Due to licensing restrictions, we cannot distribute pre-exported model assets for this model.
Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models-cn/blob/main/src/qai_hub_models/models/yoloe_seg) Python library to compile and export the model with your own:
- Custom weights (e.g., fine-tuned checkpoints)
- Custom input shapes
- Target device and runtime configurations

See our repository for [YOLOE-Segmentation on GitHub](https://github.com/qualcomm/ai-hub-models-cn/blob/main/src/qai_hub_models/models/yoloe_seg) for usage instructions.


## Model Details

**Model Type:** Model_use_case.semantic_segmentation

**Model Stats:**
- Input resolution: 640x640
- Model checkpoint: YOLOE-v8L-SEG
- Model size (float): 103 MB
- Number of parameters: 53.47M

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| YOLOE-Segmentation | ONNX | float | Snapdragon® X2 Elite | 15.738 ms | 17 - 17 MB | NPU
| YOLOE-Segmentation | ONNX | float | Snapdragon® X Elite | 31.128 ms | 88 - 88 MB | NPU
| YOLOE-Segmentation | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 24.667 ms | 3 - 286 MB | NPU
| YOLOE-Segmentation | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 66.86 ms | 17 - 345 MB | NPU
| YOLOE-Segmentation | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 55.253 ms | 9 - 17 MB | NPU
| YOLOE-Segmentation | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 31.86 ms | 0 - 221 MB | NPU
| YOLOE-Segmentation | ONNX | float | Qualcomm® QCS8450 | 66.86 ms | 17 - 345 MB | NPU
| YOLOE-Segmentation | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 47.416 ms | 13 - 20 MB | NPU
| YOLOE-Segmentation | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 31.128 ms | 88 - 88 MB | NPU
| YOLOE-Segmentation | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 18.368 ms | 13 - 232 MB | NPU
| YOLOE-Segmentation | ONNX | float | Snapdragon® 8 Elite Mobile | 18.368 ms | 13 - 232 MB | NPU
| YOLOE-Segmentation | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 12.371 ms | 12 - 249 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Snapdragon® X2 Elite | 15.565 ms | 5 - 5 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Snapdragon® X Elite | 29.2 ms | 5 - 5 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 21.739 ms | 5 - 274 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 62.692 ms | 0 - 309 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 51.952 ms | 5 - 16 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 28.147 ms | 5 - 7 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® SA8775P | 44.242 ms | 1 - 208 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® SA8650P | 44.242 ms | 1 - 208 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® SA8255P | 44.242 ms | 1 - 208 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® QCS8450 | 62.692 ms | 0 - 309 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 44.525 ms | 7 - 17 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 29.2 ms | 5 - 5 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 17.211 ms | 0 - 213 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® SA7255P | 178.394 ms | 0 - 207 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Qualcomm® SA8295P | 49.911 ms | 0 - 248 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 17.211 ms | 0 - 213 MB | NPU
| YOLOE-Segmentation | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 11.649 ms | 5 - 216 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 21.129 ms | 4 - 348 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 62.613 ms | 5 - 395 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 51.243 ms | 4 - 107 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 27.771 ms | 4 - 36 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® SA8775P | 43.578 ms | 4 - 236 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® SA8650P | 43.578 ms | 4 - 236 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® SA8255P | 43.578 ms | 4 - 236 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® QCS8450 | 62.613 ms | 5 - 395 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 43.956 ms | 4 - 106 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 16.432 ms | 4 - 241 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® SA7255P | 177.518 ms | 5 - 238 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Qualcomm® SA8295P | 48.52 ms | 5 - 277 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Snapdragon® 8 Elite Mobile | 16.432 ms | 4 - 241 MB | NPU
| YOLOE-Segmentation | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 11.328 ms | 4 - 247 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 6.034 ms | 1 - 302 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 12.967 ms | 1 - 306 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 46.38 ms | 1 - 51 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-8275 | 8.508 ms | 1 - 51 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 7.893 ms | 1 - 32 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® SA8775P | 8.092 ms | 1 - 232 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® SA8650P | 8.092 ms | 1 - 232 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® SA8255P | 8.092 ms | 1 - 232 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® QCS8450 | 12.967 ms | 1 - 306 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 7.802 ms | 1 - 51 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 148.261 ms | 1 - 248 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 13.644 ms | 1 - 229 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 4.883 ms | 1 - 230 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® SA7255P | 26.74 ms | 1 - 230 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Qualcomm® SA8295P | 14.205 ms | 1 - 233 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 4.883 ms | 1 - 230 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 3.402 ms | 0 - 242 MB | NPU
| YOLOE-Segmentation | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 13.644 ms | 1 - 229 MB | NPU

## License
* The license for the original implementation of YOLOE-Segmentation can be found
  [here](https://github.com/THU-MIG/yoloe/blob/main/LICENSE).

## References
* [YOLOE: Real-Time Seeing Anything](https://arxiv.org/abs/2503.07465)
* [Source Model Implementation](https://github.com/THU-MIG/yoloe)

## Community
* For questions or feedback, chat with our engineers in [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.cn).