Buckets:

gonduras
/

PSPNet-bucket

Files

xet

gonduras/PSPNet-bucket / README.md

gonduras

about 1 month ago

preview code

download

raw

7.3 kB

---
library_name: pytorch
license: mit
tags:
- bu_auto
- android
pipeline_tag: image-segmentation

---
![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pspnet/web-assets/model_demo.png)

# 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](https://github.com/qualcomm/ai-hub-models/blob/v0.62.2/src/qai_hub_models/models/pspnet) 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 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pspnet/releases/v0.62.2/pspnet-onnx-float.zip)
| QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pspnet/releases/v0.62.2/pspnet-qnn_dlc-float.zip)
| TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pspnet/releases/v0.62.2/pspnet-tflite-float.zip)
For more device-specific assets and performance metrics, visit **[PSPNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/pspnet)**.
### Option 2: Export with Custom Configurations
Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.62.2/src/qai_hub_models/models/pspnet) 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](https://github.com/qualcomm/ai-hub-models/blob/v0.62.2/src/qai_hub_models/models/pspnet) 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® X2 Elite | 837.924 ms | 528 - 528 MB | NPU
| PSPNet | ONNX | float | Snapdragon® X Elite | 1344.764 ms | 267 - 267 MB | NPU
| PSPNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 971.916 ms | 144 - 1989 MB | NPU
| PSPNet | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 2243.94 ms | 111 - 966 MB | NPU
| PSPNet | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 1392.889 ms | 8 - 14 MB | NPU
| PSPNet | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 1153.578 ms | 3 - 163 MB | NPU
| PSPNet | ONNX | float | Qualcomm® QCS8450 | 2243.94 ms | 111 - 966 MB | NPU
| PSPNet | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 1414.96 ms | 8 - 13 MB | NPU
| PSPNet | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 1344.764 ms | 267 - 267 MB | NPU
| PSPNet | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 645.629 ms | 205 - 1676 MB | NPU
| PSPNet | ONNX | float | Snapdragon® 8 Elite Mobile | 645.629 ms | 205 - 1676 MB | NPU
| PSPNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 719.518 ms | 242 - 1838 MB | NPU
| PSPNet | QNN_DLC | float | Snapdragon® X2 Elite | 2498.773 ms | 3 - 3 MB | NPU
| PSPNet | QNN_DLC | float | Snapdragon® X Elite | 2550.3 ms | 3 - 3 MB | NPU
| PSPNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 1844.443 ms | 3 - 1656 MB | NPU
| PSPNet | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 1591.981 ms | 1 - 851 MB | NPU
| PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 2521.737 ms | 3 - 136 MB | NPU
| PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 2503.326 ms | 0 - 26 MB | NPU
| PSPNet | QNN_DLC | float | Qualcomm® SA8775P | 2618.077 ms | 2 - 1308 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 | 1591.981 ms | 1 - 851 MB | NPU
| PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 2592.708 ms | 5 - 137 MB | NPU
| PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 2550.3 ms | 3 - 3 MB | NPU
| PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 2167.937 ms | 0 - 1311 MB | NPU
| PSPNet | QNN_DLC | float | Qualcomm® SA7255P | 5287.85 ms | 2 - 1307 MB | NPU
| PSPNet | QNN_DLC | float | Qualcomm® SA8295P | 1359.349 ms | 3 - 647 MB | NPU
| PSPNet | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 2167.937 ms | 0 - 1311 MB | NPU
| PSPNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 2327.151 ms | 0 - 1362 MB | NPU
| PSPNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2122.19 ms | 1 - 1704 MB | NPU
| PSPNet | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 1860.641 ms | 41 - 969 MB | NPU
| PSPNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 2876.544 ms | 24 - 299 MB | NPU
| PSPNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 2827.598 ms | 27 - 30 MB | NPU
| PSPNet | TFLITE | float | Qualcomm® SA8775P | 2948.943 ms | 104 - 1502 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 | 1860.641 ms | 41 - 969 MB | NPU
| PSPNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 2934.449 ms | 38 - 313 MB | NPU
| PSPNet | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 2165.949 ms | 104 - 1511 MB | NPU
| PSPNet | TFLITE | float | Qualcomm® SA7255P | 5966.001 ms | 105 - 1503 MB | NPU
| PSPNet | TFLITE | float | Qualcomm® SA8295P | 1421.718 ms | 128 - 837 MB | NPU
| PSPNet | TFLITE | float | Snapdragon® 8 Elite Mobile | 2165.949 ms | 104 - 1511 MB | NPU
| PSPNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2329.89 ms | 0 - 1451 MB | NPU

## License
* The license for the original implementation of PSPNet can be found
  [here](https://github.com/hszhao/semseg/blob/master/LICENSE).



## 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).

Xet Storage Details

Size:: 7.3 kB
Xet hash:: d9c6ffe5f30041b0a36a24a136b39ceb4a9a2b481782a13d09a6919b0b5e4830

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.

	---
	library_name: pytorch
	license: mit
	tags:
	- bu_auto
	- android
	pipeline_tag: image-segmentation

	---

	![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pspnet/web-assets/model_demo.png)

	# 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](https://github.com/qualcomm/ai-hub-models/blob/v0.62.2/src/qai_hub_models/models/pspnet) 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 2.45, ONNX Runtime 1.27.1 \| [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pspnet/releases/v0.62.2/pspnet-onnx-float.zip)
	\| QNN_DLC \| float \| Universal \| QAIRT 2.45 \| [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pspnet/releases/v0.62.2/pspnet-qnn_dlc-float.zip)
	\| TFLITE \| float \| Universal \| QAIRT 2.45 \| [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pspnet/releases/v0.62.2/pspnet-tflite-float.zip)

	For more device-specific assets and performance metrics, visit [PSPNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/pspnet).


	### Option 2: Export with Custom Configurations

	Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.62.2/src/qai_hub_models/models/pspnet) 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](https://github.com/qualcomm/ai-hub-models/blob/v0.62.2/src/qai_hub_models/models/pspnet) 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® X2 Elite \| 837.924 ms \| 528 - 528 MB \| NPU
	\| PSPNet \| ONNX \| float \| Snapdragon® X Elite \| 1344.764 ms \| 267 - 267 MB \| NPU
	\| PSPNet \| ONNX \| float \| Snapdragon® 8 Gen 3 Mobile \| 971.916 ms \| 144 - 1989 MB \| NPU
	\| PSPNet \| ONNX \| float \| Snapdragon® 8 Gen 1 Mobile \| 2243.94 ms \| 111 - 966 MB \| NPU
	\| PSPNet \| ONNX \| float \| Qualcomm® Dragonwing™ IQ-8275 \| 1392.889 ms \| 8 - 14 MB \| NPU
	\| PSPNet \| ONNX \| float \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 1153.578 ms \| 3 - 163 MB \| NPU
	\| PSPNet \| ONNX \| float \| Qualcomm® QCS8450 \| 2243.94 ms \| 111 - 966 MB \| NPU
	\| PSPNet \| ONNX \| float \| Qualcomm® Dragonwing™ IQ-9075 \| 1414.96 ms \| 8 - 13 MB \| NPU
	\| PSPNet \| ONNX \| float \| Qualcomm® Dragonwing™ IQ-X7181 \| 1344.764 ms \| 267 - 267 MB \| NPU
	\| PSPNet \| ONNX \| float \| Qualcomm® Dragonwing™ Q-8750 \| 645.629 ms \| 205 - 1676 MB \| NPU
	\| PSPNet \| ONNX \| float \| Snapdragon® 8 Elite Mobile \| 645.629 ms \| 205 - 1676 MB \| NPU
	\| PSPNet \| ONNX \| float \| Snapdragon® 8 Elite Gen 5 Mobile \| 719.518 ms \| 242 - 1838 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Snapdragon® X2 Elite \| 2498.773 ms \| 3 - 3 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Snapdragon® X Elite \| 2550.3 ms \| 3 - 3 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Snapdragon® 8 Gen 3 Mobile \| 1844.443 ms \| 3 - 1656 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Snapdragon® 8 Gen 1 Mobile \| 1591.981 ms \| 1 - 851 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Qualcomm® Dragonwing™ IQ-8275 \| 2521.737 ms \| 3 - 136 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 2503.326 ms \| 0 - 26 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Qualcomm® SA8775P \| 2618.077 ms \| 2 - 1308 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 \| 1591.981 ms \| 1 - 851 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Qualcomm® Dragonwing™ IQ-9075 \| 2592.708 ms \| 5 - 137 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Qualcomm® Dragonwing™ IQ-X7181 \| 2550.3 ms \| 3 - 3 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Qualcomm® Dragonwing™ Q-8750 \| 2167.937 ms \| 0 - 1311 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Qualcomm® SA7255P \| 5287.85 ms \| 2 - 1307 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Qualcomm® SA8295P \| 1359.349 ms \| 3 - 647 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Snapdragon® 8 Elite Mobile \| 2167.937 ms \| 0 - 1311 MB \| NPU
	\| PSPNet \| QNN_DLC \| float \| Snapdragon® 8 Elite Gen 5 Mobile \| 2327.151 ms \| 0 - 1362 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Snapdragon® 8 Gen 3 Mobile \| 2122.19 ms \| 1 - 1704 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Snapdragon® 8 Gen 1 Mobile \| 1860.641 ms \| 41 - 969 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Qualcomm® Dragonwing™ IQ-8275 \| 2876.544 ms \| 24 - 299 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 2827.598 ms \| 27 - 30 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Qualcomm® SA8775P \| 2948.943 ms \| 104 - 1502 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 \| 1860.641 ms \| 41 - 969 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Qualcomm® Dragonwing™ IQ-9075 \| 2934.449 ms \| 38 - 313 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Qualcomm® Dragonwing™ Q-8750 \| 2165.949 ms \| 104 - 1511 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Qualcomm® SA7255P \| 5966.001 ms \| 105 - 1503 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Qualcomm® SA8295P \| 1421.718 ms \| 128 - 837 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Snapdragon® 8 Elite Mobile \| 2165.949 ms \| 104 - 1511 MB \| NPU
	\| PSPNet \| TFLITE \| float \| Snapdragon® 8 Elite Gen 5 Mobile \| 2329.89 ms \| 0 - 1451 MB \| NPU

	## License
	* The license for the original implementation of PSPNet can be found
	[here](https://github.com/hszhao/semseg/blob/master/LICENSE).



	## 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).