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 2.42, ONNX Runtime 1.24.3 Download
ONNX w8a8 Universal QAIRT 2.42, ONNX Runtime 1.24.3 Download
QNN_DLC float Universal QAIRT 2.45 Download
QNN_DLC w8a8 Universal QAIRT 2.45 Download
TFLITE float Universal QAIRT 2.45 Download
TFLITE w8a8 Universal QAIRT 2.45 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:

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

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
PSPNet ONNX float Snapdragon® 8 Elite Gen 5 Mobile 321.368 ms 239 - 1007 MB NPU
PSPNet ONNX float Snapdragon® 8 Elite Mobile 335.615 ms 275 - 995 MB NPU
PSPNet ONNX float Snapdragon® X2 Elite 429.747 ms 265 - 265 MB NPU
PSPNet ONNX float Snapdragon® X Elite 696.139 ms 529 - 529 MB NPU
PSPNet ONNX float Snapdragon® X Elite 696.139 ms 529 - 529 MB NPU
PSPNet ONNX float Snapdragon® 8 Gen 3 Mobile 508.053 ms 147 - 1063 MB NPU
PSPNet ONNX float Qualcomm® QCS8550 (Proxy) 604.294 ms 0 - 322 MB NPU
PSPNet ONNX float Qualcomm® QCS9075 1755.826 ms 80 - 86 MB NPU
PSPNet ONNX float Snapdragon® 8 Elite For Galaxy Mobile 335.615 ms 275 - 995 MB NPU
PSPNet ONNX w8a8 Snapdragon® 8 Elite Gen 5 Mobile 16.879 ms 52 - 320 MB NPU
PSPNet ONNX w8a8 Snapdragon® 8 Elite Mobile 19.121 ms 69 - 338 MB NPU
PSPNet ONNX w8a8 Snapdragon® X2 Elite 15.19 ms 131 - 131 MB NPU
PSPNet ONNX w8a8 Snapdragon® X Elite 28.774 ms 133 - 133 MB NPU
PSPNet ONNX w8a8 Snapdragon® X Elite 28.774 ms 133 - 133 MB NPU
PSPNet ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 22.145 ms 69 - 490 MB NPU
PSPNet ONNX w8a8 Qualcomm® QCS6490 3312.591 ms 205 - 284 MB CPU
PSPNet ONNX w8a8 Qualcomm® QCS8550 (Proxy) 30.284 ms 64 - 278 MB NPU
PSPNet ONNX w8a8 Qualcomm® QCS9075 34.17 ms 70 - 73 MB NPU
PSPNet ONNX w8a8 Qualcomm® QCM6690 3291.834 ms 55 - 66 MB CPU
PSPNet ONNX w8a8 Snapdragon® 8 Elite For Galaxy Mobile 19.121 ms 69 - 338 MB NPU
PSPNet ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 2621.081 ms 57 - 69 MB CPU
PSPNet ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 2621.081 ms 57 - 69 MB CPU
PSPNet QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 270.932 ms 3 - 748 MB NPU
PSPNet QNN_DLC float Snapdragon® 8 Elite Mobile 301.987 ms 0 - 725 MB NPU
PSPNet QNN_DLC float Snapdragon® X2 Elite 229.067 ms 3 - 3 MB NPU
PSPNet QNN_DLC float Snapdragon® X Elite 530.829 ms 3 - 3 MB NPU
PSPNet QNN_DLC float Snapdragon® X Elite 530.829 ms 3 - 3 MB NPU
PSPNet QNN_DLC float Snapdragon® 8 Gen 3 Mobile 398.384 ms 1 - 950 MB NPU
PSPNet QNN_DLC float Qualcomm® QCS8275 (Proxy) 1332.245 ms 3 - 726 MB NPU
PSPNet QNN_DLC float Qualcomm® QCS8550 (Proxy) 592.037 ms 3 - 645 MB NPU
PSPNet QNN_DLC float Qualcomm® QCS9075 1757.735 ms 3 - 135 MB NPU
PSPNet QNN_DLC float Qualcomm® QCS8450 (Proxy) 1701.421 ms 0 - 439 MB NPU
PSPNet QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 301.987 ms 0 - 725 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 Mobile 15.232 ms 1 - 310 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 8 Elite Mobile 18.275 ms 1 - 231 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® X2 Elite 14.335 ms 1 - 1 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® X Elite 26.154 ms 1 - 1 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® X Elite 26.154 ms 1 - 1 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 20.652 ms 0 - 355 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® QCS8275 (Proxy) 63.816 ms 0 - 259 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® QCS8550 (Proxy) 28.589 ms 1 - 3 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® QCS9075 32.103 ms 1 - 35 MB NPU
PSPNet QNN_DLC w8a8 Qualcomm® QCS8450 (Proxy) 37.916 ms 1 - 351 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 8 Elite For Galaxy Mobile 18.275 ms 1 - 231 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 91.147 ms 1 - 446 MB NPU
PSPNet QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 91.147 ms 1 - 446 MB NPU
PSPNet TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 271.382 ms 128 - 912 MB NPU
PSPNet TFLITE float Snapdragon® 8 Elite Mobile 302.326 ms 127 - 904 MB NPU
PSPNet TFLITE float Snapdragon® 8 Gen 3 Mobile 405.485 ms 127 - 1124 MB NPU
PSPNet TFLITE float Qualcomm® QCS8275 (Proxy) 1328.645 ms 115 - 888 MB NPU
PSPNet TFLITE float Qualcomm® QCS8550 (Proxy) 592.371 ms 120 - 124 MB NPU
PSPNet TFLITE float Qualcomm® QCS9075 1764.985 ms 0 - 271 MB NPU
PSPNet TFLITE float Qualcomm® QCS8450 (Proxy) 1509.154 ms 91 - 641 MB NPU
PSPNet TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 302.326 ms 127 - 904 MB NPU
PSPNet TFLITE w8a8 Snapdragon® 8 Elite Gen 5 Mobile 15.015 ms 32 - 311 MB NPU
PSPNet TFLITE w8a8 Snapdragon® 8 Elite Mobile 18.149 ms 32 - 270 MB NPU
PSPNet TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 20.811 ms 32 - 389 MB NPU
PSPNet TFLITE w8a8 Qualcomm® QCS8275 (Proxy) 64.182 ms 32 - 282 MB NPU
PSPNet TFLITE w8a8 Qualcomm® QCS8550 (Proxy) 28.665 ms 32 - 34 MB NPU
PSPNet TFLITE w8a8 Qualcomm® QCS9075 41.172 ms 32 - 132 MB NPU
PSPNet TFLITE w8a8 Qualcomm® QCS8450 (Proxy) 36.814 ms 32 - 387 MB NPU
PSPNet TFLITE w8a8 Snapdragon® 8 Elite For Galaxy Mobile 18.149 ms 32 - 270 MB NPU
PSPNet TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 91.251 ms 32 - 441 MB NPU
PSPNet TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 91.251 ms 32 - 441 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
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for qualcomm/PSPNet

Finetunes
1 model