Hugging Face's logo

qualcomm
/
PSPNet

Image Segmentation
PyTorch
bu_auto
android
Model card Files
xet
 Community

PSPNet: Optimized for Qualcomm Devices

PSPNet (Pyramid Scene Parsing Network) is a semantic segmentation model that captures global context information by applying pyramid pooling modules. It is designed to improve scene understanding by aggregating contextual features at multiple scales.

This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.

Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up 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 2.50, ONNX Runtime 1.27.1 Download
ONNX w8a8 Universal QAIRT 2.50, ONNX Runtime 1.27.1 Download
QNN_DLC float Universal QAIRT 2.50 Download
QNN_DLC w8a8 Universal QAIRT 2.50 Download
TFLITE float Universal QAIRT 2.50 Download

For more device-specific assets and performance metrics, visit PSPNet on Qualcomm® AI Hub.

Option 2: Export with Custom Configurations

Use the Qualcomm® AI Hub Models 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 PSPNet on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

  • Input resolution: 1x3x473x473
  • Model checkpoint: pspnet101_ade20k.pth
  • Model size (float): 251 MB
  • Number of parameters: 65.7M

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
PSPNet ONNX float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 534.749 ms 151 - 1806 MB NPU
PSPNet ONNX float Snapdragon® 8 Elite For Galaxy Mobile 584.12 ms 103 - 1747 MB NPU
PSPNet ONNX float Snapdragon® X2 Elite 530.022 ms 529 - 529 MB NPU
PSPNet ONNX float Snapdragon® X Elite 774.488 ms 532 - 532 MB NPU
PSPNet ONNX float Snapdragon® 8 Gen 3 Mobile 593.617 ms 142 - 2936 MB NPU
PSPNet ONNX float Snapdragon® 8 Gen 1 Mobile 1450.965 ms 14 - 887 MB NPU
PSPNet ONNX float Qualcomm® Dragonwing™ IQ-8275 893.451 ms 8 - 15 MB NPU
PSPNet ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 778.267 ms 218 - 221 MB NPU
PSPNet ONNX float Qualcomm® QCS8450 1450.965 ms 14 - 887 MB NPU
PSPNet ONNX float Qualcomm® Dragonwing™ IQ-9075 860.587 ms 8 - 14 MB NPU
PSPNet ONNX float Qualcomm® Dragonwing™ IQ-X7181 774.488 ms 532 - 532 MB NPU
PSPNet ONNX float Qualcomm® Dragonwing™ Q-8750 584.12 ms 103 - 1747 MB NPU
PSPNet ONNX w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 37.894 ms 69 - 561 MB NPU
PSPNet ONNX w8a8 Snapdragon® 8 Elite For Galaxy Mobile 43.809 ms 70 - 444 MB NPU
PSPNet ONNX w8a8 Snapdragon® X2 Elite 35.351 ms 132 - 132 MB NPU
PSPNet ONNX w8a8 Snapdragon® X Elite 65.604 ms 133 - 133 MB NPU
PSPNet ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 51.517 ms 72 - 689 MB NPU
PSPNet ONNX w8a8 Snapdragon® 8 Gen 1 Mobile 81.97 ms 0 - 557 MB NPU
PSPNet ONNX w8a8 Qualcomm® Dragonwing™ IQ-8275 68.476 ms 68 - 73 MB NPU
PSPNet ONNX w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 68.707 ms 0 - 83 MB NPU
PSPNet ONNX w8a8 Qualcomm® QCS8450 81.97 ms 0 - 557 MB NPU
PSPNet ONNX w8a8 Qualcomm® Dragonwing™ IQ-9075 67.803 ms 71 - 75 MB NPU
PSPNet ONNX w8a8 Qualcomm® Dragonwing™ IQ-X7181 65.604 ms 133 - 133 MB NPU
PSPNet ONNX w8a8 Qualcomm® Dragonwing™ Q-7790 147.868 ms 71 - 572 MB NPU
PSPNet ONNX w8a8 Qualcomm® Dragonwing™ Q-8750 43.809 ms 70 - 444 MB NPU
PSPNet ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 147.868 ms 71 - 572 MB NPU
PSPNet QNN_DLC float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 546.312 ms 0 - 1652 MB NPU
PSPNet QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 571.78 ms 3 - 1594 MB NPU
PSPNet QNN_DLC float Snapdragon® X2 Elite 545.219 ms 3 - 3 MB NPU
PSPNet QNN_DLC float Snapdragon® X Elite 787.972 ms 3 - 3 MB NPU
PSPNet QNN_DLC float Snapdragon® 8 Gen 3 Mobile 621.508 ms 3 - 2801 MB NPU
PSPNet QNN_DLC float Snapdragon® 8 Gen 1 Mobile 1073.257 ms 0 - 703 MB NPU
PSPNet QNN_DLC float Qualcomm® Dragonwing™ IQ-8275 913.255 ms 3 - 136 MB NPU
PSPNet QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 797.908 ms 3 - 1360 MB NPU
PSPNet QNN_DLC float Qualcomm® SA8650P 2618.077 ms 2 - 1308 MB NPU
PSPNet QNN_DLC float Qualcomm® SA8255P 2618.077 ms 2 - 1308 MB NPU
PSPNet QNN_DLC float Qualcomm® QCS8450 1073.257 ms 0 - 703 MB NPU
PSPNet QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 868.583 ms 5 - 137 MB NPU
PSPNet QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 787.972 ms 3 - 3 MB NPU
PSPNet QNN_DLC float Qualcomm® Dragonwing™ Q-8750 571.78 ms 3 - 1594 MB NPU
PSPNet QNN_DLC float Qualcomm® SA8295P 966.467 ms 10 - 566 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 41.077 ms 1 - 571 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 8 Elite For Galaxy Mobile 49.024 ms 0 - 376 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® X2 Elite 37.337 ms 1 - 1 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® X Elite 70.915 ms 1 - 1 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 58.01 ms 1 - 596 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 8 Gen 1 Mobile 92.296 ms 1 - 550 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-8275 75.218 ms 1 - 36 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 75.224 ms 1 - 548 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® QCS8450 92.296 ms 1 - 550 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-9075 73.595 ms 1 - 35 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-X7181 70.915 ms 1 - 1 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-7790 146.11 ms 1 - 486 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-8750 49.024 ms 0 - 376 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® SA8295P 98.682 ms 1 - 437 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 146.11 ms 1 - 486 MB NPU
PSPNet TFLITE float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 554.788 ms 111 - 1804 MB NPU
PSPNet TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 584.683 ms 127 - 1760 MB NPU
PSPNet TFLITE float Snapdragon® 8 Gen 3 Mobile 605.29 ms 134 - 3020 MB NPU
PSPNet TFLITE float Snapdragon® 8 Gen 1 Mobile 1445.795 ms 83 - 1008 MB NPU
PSPNet TFLITE float Qualcomm® Dragonwing™ IQ-8275 906.913 ms 0 - 279 MB NPU
PSPNet TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 779.948 ms 128 - 131 MB NPU
PSPNet TFLITE float Qualcomm® SA8650P 2948.943 ms 104 - 1502 MB NPU
PSPNet TFLITE float Qualcomm® SA8255P 2948.943 ms 104 - 1502 MB NPU
PSPNet TFLITE float Qualcomm® QCS8450 1445.795 ms 83 - 1008 MB NPU
PSPNet TFLITE float Qualcomm® Dragonwing™ IQ-9075 871.696 ms 0 - 278 MB NPU
PSPNet TFLITE float Qualcomm® Dragonwing™ Q-8750 584.683 ms 127 - 1760 MB NPU
PSPNet TFLITE float Qualcomm® SA7255P 11763.129 ms 516 - 526 MB CPU
PSPNet TFLITE float Qualcomm® SA8295P 890.415 ms 93 - 807 MB NPU

License

  • The license for the original implementation of PSPNet can be found here.

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
Image Segmentation
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for qualcomm/PSPNet

Finetunes
1 model