--- library_name: pytorch license: other tags: - real_time - android pipeline_tag: image-segmentation --- ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/yolo26_seg/web-assets/model_demo.png) # YOLO26-Segmentation: Optimized for Qualcomm Devices Ultralytics YOLO26 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 YOLO26-Segmentation found [here](https://github.com/ultralytics/ultralytics/tree/main/ultralytics/models/yolo/segment). 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/blob/main/src/qai_hub_models/models/yolo26_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.com) 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/blob/main/src/qai_hub_models/models/yolo26_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 [YOLO26-Segmentation on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/yolo26_seg) for usage instructions. ## Model Details **Model Type:** Model_use_case.semantic_segmentation **Model Stats:** - Model checkpoint: YOLO26N-Seg - Input resolution: 640x640 - Number of output classes: 80 - Number of parameters: 2.7M - Model size (float): 9.29 MB - Model size (w8a16): 5.25 MB ## Performance Summary | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |---|---|---|---|---|---|--- | YOLO26-Segmentation | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 4.088 ms | 14 - 189 MB | NPU | YOLO26-Segmentation | ONNX | float | Snapdragon® X2 Elite | 3.931 ms | 16 - 16 MB | NPU | YOLO26-Segmentation | ONNX | float | Snapdragon® X Elite | 7.331 ms | 17 - 17 MB | NPU | YOLO26-Segmentation | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 4.79 ms | 16 - 229 MB | NPU | YOLO26-Segmentation | ONNX | float | Qualcomm® QCS8550 (Proxy) | 6.891 ms | 11 - 19 MB | NPU | YOLO26-Segmentation | ONNX | float | Qualcomm® QCS9075 | 9.949 ms | 13 - 15 MB | NPU | YOLO26-Segmentation | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.562 ms | 11 - 185 MB | NPU | YOLO26-Segmentation | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.661 ms | 0 - 94 MB | NPU | YOLO26-Segmentation | ONNX | w8a16 | Snapdragon® X2 Elite | 2.88 ms | 6 - 6 MB | NPU | YOLO26-Segmentation | ONNX | w8a16 | Snapdragon® X Elite | 6.852 ms | 7 - 7 MB | NPU | YOLO26-Segmentation | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.896 ms | 8 - 248 MB | NPU | YOLO26-Segmentation | ONNX | w8a16 | Qualcomm® QCS6490 | 458.899 ms | 174 - 180 MB | CPU | YOLO26-Segmentation | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 6.276 ms | 8 - 13 MB | NPU | YOLO26-Segmentation | ONNX | w8a16 | Qualcomm® QCS9075 | 7.135 ms | 8 - 11 MB | NPU | YOLO26-Segmentation | ONNX | w8a16 | Qualcomm® QCM6690 | 216.63 ms | 116 - 127 MB | CPU | YOLO26-Segmentation | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.982 ms | 2 - 89 MB | NPU | YOLO26-Segmentation | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 194.553 ms | 171 - 181 MB | CPU | YOLO26-Segmentation | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.249 ms | 5 - 169 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Snapdragon® X2 Elite | 3.087 ms | 5 - 5 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Snapdragon® X Elite | 5.489 ms | 5 - 5 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 3.678 ms | 5 - 189 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 16.395 ms | 1 - 159 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 5.029 ms | 5 - 6 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Qualcomm® SA8775P | 6.206 ms | 0 - 160 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Qualcomm® QCS9075 | 7.354 ms | 5 - 15 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 10.566 ms | 3 - 207 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Qualcomm® SA7255P | 16.395 ms | 1 - 159 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Qualcomm® SA8295P | 10.858 ms | 0 - 175 MB | NPU | YOLO26-Segmentation | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.781 ms | 0 - 166 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.315 ms | 2 - 194 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 2.765 ms | 2 - 2 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Snapdragon® X Elite | 5.769 ms | 2 - 2 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.595 ms | 2 - 210 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 9.697 ms | 0 - 186 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 5.271 ms | 2 - 4 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Qualcomm® SA8775P | 5.982 ms | 1 - 186 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 6.002 ms | 2 - 8 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 26.872 ms | 2 - 187 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Qualcomm® SA7255P | 9.697 ms | 0 - 186 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.645 ms | 0 - 190 MB | NPU | YOLO26-Segmentation | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 5.27 ms | 2 - 195 MB | NPU | YOLO26-Segmentation | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.633 ms | 0 - 167 MB | NPU | YOLO26-Segmentation | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 4.15 ms | 3 - 192 MB | NPU | YOLO26-Segmentation | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 17.306 ms | 4 - 165 MB | NPU | YOLO26-Segmentation | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 5.489 ms | 4 - 6 MB | NPU | YOLO26-Segmentation | TFLITE | float | Qualcomm® SA8775P | 7.025 ms | 4 - 170 MB | NPU | YOLO26-Segmentation | TFLITE | float | Qualcomm® QCS9075 | 7.934 ms | 4 - 22 MB | NPU | YOLO26-Segmentation | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 11.385 ms | 4 - 209 MB | NPU | YOLO26-Segmentation | TFLITE | float | Qualcomm® SA7255P | 17.306 ms | 4 - 165 MB | NPU | YOLO26-Segmentation | TFLITE | float | Qualcomm® SA8295P | 11.407 ms | 4 - 182 MB | NPU | YOLO26-Segmentation | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.195 ms | 0 - 161 MB | NPU ## License * The license for the original implementation of YOLO26-Segmentation can be found [here](https://github.com/ultralytics/ultralytics/blob/main/LICENSE). ## References * [Ultralytics YOLO26 Docs: Instance Segmentation](https://docs.ultralytics.com/tasks/segment/) * [Source Model Implementation](https://github.com/ultralytics/ultralytics/tree/main/ultralytics/models/yolo/segment) ## Community * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI. * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).