Hugging Face's logo

qualcomm
/
SINet

Image Segmentation
PyTorch
android
Model card Files
xet
 Community
SINet / README.md
qaihm-bot's picture
qaihm-bot
v0.54.0
1d0b8f9 verified
preview code
|
raw
history blame contribute delete
13.2 kB
---
library_name: pytorch
license: other
tags:
- android
pipeline_tag: image-segmentation
---
![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/web-assets/model_demo.png)
# SINet: Optimized for Qualcomm Devices
SINet is a machine learning model that is designed to segment people from close-up portrait images in real time.
This is based on the implementation of SINet found [here](https://github.com/clovaai/ext_portrait_segmentation).
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/sinet) 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
There are two ways to deploy this model on your device:
### Option 1: Download Pre-Exported Models
Below are pre-exported model assets ready for deployment.
| Runtime | Precision | Chipset | SDK Versions | Download |
|---|---|---|---|---|
| ONNX | float | Universal | QAIRT 2.42, ONNX Runtime 1.24.3 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/releases/v0.54.0/sinet-onnx-float.zip)
| ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.3 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/releases/v0.54.0/sinet-onnx-w8a16.zip)
| ONNX | w8a8 | Universal | QAIRT 2.42, ONNX Runtime 1.24.3 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/releases/v0.54.0/sinet-onnx-w8a8.zip)
| QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/releases/v0.54.0/sinet-qnn_dlc-float.zip)
| QNN_DLC | w8a16 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/releases/v0.54.0/sinet-qnn_dlc-w8a16.zip)
| QNN_DLC | w8a8 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/releases/v0.54.0/sinet-qnn_dlc-w8a8.zip)
| TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/releases/v0.54.0/sinet-tflite-float.zip)
| TFLITE | w8a8 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sinet/releases/v0.54.0/sinet-tflite-w8a8.zip)
For more device-specific assets and performance metrics, visit **[SINet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/sinet)**.
### Option 2: Export with Custom Configurations
Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/sinet) 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
This option is ideal if you need to customize the model beyond the default configuration provided here.
See our repository for [SINet on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/sinet) for usage instructions.
## Model Details
**Model Type:** Model_use_case.semantic_segmentation
**Model Stats:**
- Model checkpoint: SINet.pth
- Input resolution: 224x224
- Number of output classes: 2 (foreground / background)
- Number of parameters: 91.9K
- Model size (float): 415 KB
- Model size (w8a8): 241 KB
## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| SINet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.706 ms | 0 - 40 MB | NPU
| SINet | ONNX | float | Snapdragon® 8 Elite Mobile | 0.855 ms | 0 - 30 MB | NPU
| SINet | ONNX | float | Snapdragon® X2 Elite | 0.766 ms | 0 - 0 MB | NPU
| SINet | ONNX | float | Snapdragon® X Elite | 1.927 ms | 2 - 2 MB | NPU
| SINet | ONNX | float | Snapdragon® X Elite | 1.927 ms | 2 - 2 MB | NPU
| SINet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 1.109 ms | 0 - 46 MB | NPU
| SINet | ONNX | float | Qualcomm® QCS8550 (Proxy) | 1.692 ms | 1 - 3 MB | NPU
| SINet | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.855 ms | 0 - 30 MB | NPU
| SINet | ONNX | float | Qualcomm® QCS9075 | 2.136 ms | 2 - 5 MB | NPU
| SINet | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 0.891 ms | 0 - 46 MB | NPU
| SINet | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 1.038 ms | 0 - 46 MB | NPU
| SINet | ONNX | w8a16 | Snapdragon® X2 Elite | 1.348 ms | 0 - 0 MB | NPU
| SINet | ONNX | w8a16 | Snapdragon® X Elite | 2.247 ms | 1 - 1 MB | NPU
| SINet | ONNX | w8a16 | Snapdragon® X Elite | 2.247 ms | 1 - 1 MB | NPU
| SINet | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 1.307 ms | 0 - 53 MB | NPU
| SINet | ONNX | w8a16 | Qualcomm® QCS6490 | 50.956 ms | 12 - 15 MB | CPU
| SINet | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 1.962 ms | 0 - 4 MB | NPU
| SINet | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 13.709 ms | 20 - 29 MB | CPU
| SINet | ONNX | w8a16 | Qualcomm® QCS9075 | 2.243 ms | 0 - 3 MB | NPU
| SINet | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 1.038 ms | 0 - 46 MB | NPU
| SINet | ONNX | w8a16 | Qualcomm® QCM6690 | 19.873 ms | 21 - 30 MB | CPU
| SINet | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 13.709 ms | 20 - 29 MB | CPU
| SINet | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 3.391 ms | 0 - 37 MB | NPU
| SINet | ONNX | w8a8 | Snapdragon® 8 Elite Mobile | 3.432 ms | 0 - 35 MB | NPU
| SINet | ONNX | w8a8 | Snapdragon® X2 Elite | 3.414 ms | 7 - 7 MB | NPU
| SINet | ONNX | w8a8 | Snapdragon® X Elite | 43.853 ms | 7 - 7 MB | NPU
| SINet | ONNX | w8a8 | Snapdragon® X Elite | 43.853 ms | 7 - 7 MB | NPU
| SINet | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 4.169 ms | 6 - 60 MB | NPU
| SINet | ONNX | w8a8 | Qualcomm® QCS6490 | 38.056 ms | 7 - 11 MB | CPU
| SINet | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 11.255 ms | 7 - 16 MB | CPU
| SINet | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 3.432 ms | 0 - 35 MB | NPU
| SINet | ONNX | w8a8 | Qualcomm® QCM6690 | 15.676 ms | 7 - 15 MB | CPU
| SINet | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 11.255 ms | 7 - 16 MB | CPU
| SINet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.124 ms | 1 - 38 MB | NPU
| SINet | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 1.808 ms | 0 - 37 MB | NPU
| SINet | QNN_DLC | float | Snapdragon® X2 Elite | 1.422 ms | 1 - 1 MB | NPU
| SINet | QNN_DLC | float | Snapdragon® X Elite | 3.821 ms | 1 - 1 MB | NPU
| SINet | QNN_DLC | float | Snapdragon® X Elite | 3.821 ms | 1 - 1 MB | NPU
| SINet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 2.109 ms | 0 - 47 MB | NPU
| SINet | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 3.43 ms | 1 - 2 MB | NPU
| SINet | QNN_DLC | float | Qualcomm® SA8775P | 3.927 ms | 1 - 36 MB | NPU
| SINet | QNN_DLC | float | Qualcomm® SA8775P | 3.927 ms | 1 - 36 MB | NPU
| SINet | QNN_DLC | float | Qualcomm® SA8775P | 3.927 ms | 1 - 36 MB | NPU
| SINet | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 3.76 ms | 0 - 52 MB | NPU
| SINet | QNN_DLC | float | Qualcomm® SA7255P | 6.652 ms | 1 - 34 MB | NPU
| SINet | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.808 ms | 0 - 37 MB | NPU
| SINet | QNN_DLC | float | Qualcomm® SA8295P | 4.656 ms | 0 - 35 MB | NPU
| SINet | QNN_DLC | float | Qualcomm® QCS9075 | 3.829 ms | 3 - 6 MB | NPU
| SINet | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 3.2 ms | 0 - 42 MB | NPU
| SINet | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 3.437 ms | 0 - 41 MB | NPU
| SINet | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 3.526 ms | 0 - 0 MB | NPU
| SINet | QNN_DLC | w8a16 | Snapdragon® X Elite | 7.09 ms | 0 - 0 MB | NPU
| SINet | QNN_DLC | w8a16 | Snapdragon® X Elite | 7.09 ms | 0 - 0 MB | NPU
| SINet | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 4.072 ms | 0 - 49 MB | NPU
| SINet | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 6.579 ms | 0 - 34 MB | NPU
| SINet | QNN_DLC | w8a16 | Qualcomm® SA8775P | 7.052 ms | 0 - 40 MB | NPU
| SINet | QNN_DLC | w8a16 | Qualcomm® SA8775P | 7.052 ms | 0 - 40 MB | NPU
| SINet | QNN_DLC | w8a16 | Qualcomm® SA8775P | 7.052 ms | 0 - 40 MB | NPU
| SINet | QNN_DLC | w8a16 | Qualcomm® QCS8450 (Proxy) | 7.173 ms | 0 - 52 MB | NPU
| SINet | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 6.825 ms | 2 - 5 MB | NPU
| SINet | QNN_DLC | w8a16 | Qualcomm® SA7255P | 10.812 ms | 0 - 37 MB | NPU
| SINet | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 3.437 ms | 0 - 41 MB | NPU
| SINet | QNN_DLC | w8a16 | Qualcomm® SA8295P | 8.169 ms | 0 - 38 MB | NPU
| SINet | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 2.137 ms | 0 - 40 MB | NPU
| SINet | QNN_DLC | w8a8 | Snapdragon® 8 Elite Mobile | 2.355 ms | 0 - 36 MB | NPU
| SINet | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 2.324 ms | 0 - 0 MB | NPU
| SINet | QNN_DLC | w8a8 | Snapdragon® X Elite | 4.911 ms | 0 - 0 MB | NPU
| SINet | QNN_DLC | w8a8 | Snapdragon® X Elite | 4.911 ms | 0 - 0 MB | NPU
| SINet | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 3.01 ms | 0 - 47 MB | NPU
| SINet | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 4.497 ms | 0 - 10 MB | NPU
| SINet | QNN_DLC | w8a8 | Qualcomm® SA8775P | 4.838 ms | 0 - 39 MB | NPU
| SINet | QNN_DLC | w8a8 | Qualcomm® SA8775P | 4.838 ms | 0 - 39 MB | NPU
| SINet | QNN_DLC | w8a8 | Qualcomm® SA8775P | 4.838 ms | 0 - 39 MB | NPU
| SINet | QNN_DLC | w8a8 | Qualcomm® QCS8450 (Proxy) | 4.631 ms | 0 - 49 MB | NPU
| SINet | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 2.355 ms | 0 - 36 MB | NPU
| SINet | QNN_DLC | w8a8 | Qualcomm® SA7255P | 8.014 ms | 0 - 36 MB | NPU
| SINet | QNN_DLC | w8a8 | Qualcomm® SA8295P | 5.22 ms | 0 - 35 MB | NPU
| SINet | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 4.443 ms | 0 - 2 MB | NPU
| SINet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.131 ms | 0 - 39 MB | NPU
| SINet | TFLITE | float | Snapdragon® 8 Elite Mobile | 1.806 ms | 0 - 36 MB | NPU
| SINet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2.112 ms | 0 - 45 MB | NPU
| SINet | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 3.391 ms | 0 - 2 MB | NPU
| SINet | TFLITE | float | Qualcomm® SA8775P | 3.921 ms | 0 - 36 MB | NPU
| SINet | TFLITE | float | Qualcomm® SA8775P | 3.921 ms | 0 - 36 MB | NPU
| SINet | TFLITE | float | Qualcomm® SA8775P | 3.921 ms | 0 - 36 MB | NPU
| SINet | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 3.79 ms | 0 - 50 MB | NPU
| SINet | TFLITE | float | Qualcomm® SA7255P | 6.668 ms | 1 - 34 MB | NPU
| SINet | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.806 ms | 0 - 36 MB | NPU
| SINet | TFLITE | float | Qualcomm® SA8295P | 4.66 ms | 1 - 35 MB | NPU
| SINet | TFLITE | float | Qualcomm® QCS9075 | 3.845 ms | 1 - 4 MB | NPU
| SINet | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 2.119 ms | 0 - 39 MB | NPU
| SINet | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 2.378 ms | 0 - 34 MB | NPU
| SINet | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 2.787 ms | 0 - 45 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® QCS6490 | 27.707 ms | 0 - 13 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 4.163 ms | 0 - 30 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® SA8775P | 4.561 ms | 0 - 39 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® SA8775P | 4.561 ms | 0 - 39 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® SA8775P | 4.561 ms | 0 - 39 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 4.282 ms | 0 - 47 MB | NPU
| SINet | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 8.575 ms | 0 - 30 MB | NPU
| SINet | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 2.378 ms | 0 - 34 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® SA7255P | 7.489 ms | 0 - 35 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® QCM6690 | 17.291 ms | 0 - 30 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® SA8295P | 5.028 ms | 0 - 34 MB | NPU
| SINet | TFLITE | w8a8 | Qualcomm® QCS9075 | 4.393 ms | 0 - 3 MB | NPU
| SINet | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 8.575 ms | 0 - 30 MB | NPU
## License
* The license for the original implementation of SINet can be found
 [here](https://github.com/clovaai/ext_portrait_segmentation/blob/master/LICENSE).
## References
* [SINet: Extreme Lightweight Portrait Segmentation Networks with Spatial Squeeze Modules and Information Blocking Decoder](https://arxiv.org/abs/1911.09099)
* [Source Model Implementation](https://github.com/clovaai/ext_portrait_segmentation)
## 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).