ResNet101: Optimized for Qualcomm Devices

ResNet101 is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.

This is based on the implementation of ResNet101 found here. 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 2.45, ONNX Runtime 1.25.0	Download
ONNX	w8a8	Universal	QAIRT 2.45, ONNX Runtime 1.25.0	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 ResNet101 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 ResNet101 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

• Model checkpoint: Imagenet
• Input resolution: 224x224
• Number of parameters: 44.5M
• Model size (float): 170 MB
• Model size (w8a8): 43.9 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
ResNet101	ONNX	float	Snapdragon® X2 Elite	1.622 ms	180 - 180 MB	NPU
ResNet101	ONNX	float	Snapdragon® X Elite	3.059 ms	148 - 148 MB	NPU
ResNet101	ONNX	float	Snapdragon® 8 Gen 3 Mobile	2.263 ms	1 - 117 MB	NPU
ResNet101	ONNX	float	Snapdragon® 8 Gen 1 Mobile	5.396 ms	1 - 86 MB	NPU
ResNet101	ONNX	float	Qualcomm® QCS8550 (Proxy)	3.066 ms	0 - 99 MB	NPU
ResNet101	ONNX	float	Qualcomm® QCS8450	5.396 ms	1 - 86 MB	NPU
ResNet101	ONNX	float	Snapdragon® 8 Elite Mobile	1.873 ms	0 - 68 MB	NPU
ResNet101	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	1.588 ms	1 - 65 MB	NPU
ResNet101	ONNX	float	Qualcomm® QCS9075	5.144 ms	1 - 46 MB	NPU
ResNet101	ONNX	float	Qualcomm® QCS8750	1.873 ms	0 - 68 MB	NPU
ResNet101	ONNX	float	Qualcomm® QCS7181	3.059 ms	148 - 148 MB	NPU
ResNet101	ONNX	w8a8	Snapdragon® X2 Elite	0.564 ms	213 - 213 MB	NPU
ResNet101	ONNX	w8a8	Snapdragon® X Elite	1.143 ms	181 - 181 MB	NPU
ResNet101	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.97 ms	0 - 133 MB	NPU
ResNet101	ONNX	w8a8	Snapdragon® 8 Gen 1 Mobile	1.8 ms	0 - 135 MB	NPU
ResNet101	ONNX	w8a8	Qualcomm® QCS6490	4.574 ms	0 - 45 MB	NPU
ResNet101	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	1.282 ms	0 - 5 MB	NPU
ResNet101	ONNX	w8a8	Qualcomm® QCS8450	1.8 ms	0 - 135 MB	NPU
ResNet101	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.814 ms	0 - 81 MB	NPU
ResNet101	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	1.759 ms	0 - 87 MB	NPU
ResNet101	ONNX	w8a8	Qualcomm® QCM6690	11.775 ms	0 - 200 MB	NPU
ResNet101	ONNX	w8a8	Qualcomm® QCS9075	1.338 ms	0 - 45 MB	NPU
ResNet101	ONNX	w8a8	Snapdragon® 8 Elite Mobile	0.835 ms	0 - 80 MB	NPU
ResNet101	ONNX	w8a8	Qualcomm® QCS7790	1.759 ms	0 - 87 MB	NPU
ResNet101	ONNX	w8a8	Qualcomm® QCS8750	0.835 ms	0 - 80 MB	NPU
ResNet101	ONNX	w8a8	Qualcomm® QCS7181	1.143 ms	181 - 181 MB	NPU
ResNet101	QNN_DLC	float	Snapdragon® X2 Elite	1.968 ms	1 - 1 MB	NPU
ResNet101	QNN_DLC	float	Snapdragon® X Elite	3.516 ms	1 - 1 MB	NPU
ResNet101	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	2.425 ms	0 - 121 MB	NPU
ResNet101	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	5.915 ms	0 - 91 MB	NPU
ResNet101	QNN_DLC	float	Qualcomm® QCS8275	18.348 ms	1 - 66 MB	NPU
ResNet101	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	3.303 ms	1 - 3 MB	NPU
ResNet101	QNN_DLC	float	Qualcomm® QCS8450	5.915 ms	0 - 91 MB	NPU
ResNet101	QNN_DLC	float	Snapdragon® 8 Elite Mobile	1.95 ms	0 - 66 MB	NPU
ResNet101	QNN_DLC	float	Qualcomm® SA8295P	5.634 ms	1 - 41 MB	NPU
ResNet101	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.646 ms	1 - 70 MB	NPU
ResNet101	QNN_DLC	float	Qualcomm® SA7255P	18.348 ms	1 - 66 MB	NPU
ResNet101	QNN_DLC	float	Qualcomm® QCS9075	5.305 ms	1 - 3 MB	NPU
ResNet101	QNN_DLC	float	Qualcomm® QCS8750	1.95 ms	0 - 66 MB	NPU
ResNet101	QNN_DLC	float	Qualcomm® QCS7181	3.516 ms	1 - 1 MB	NPU
ResNet101	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.648 ms	0 - 0 MB	NPU
ResNet101	QNN_DLC	w8a8	Snapdragon® X Elite	1.258 ms	1 - 1 MB	NPU
ResNet101	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.981 ms	0 - 127 MB	NPU
ResNet101	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	1.801 ms	0 - 128 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCS6490	4.551 ms	2 - 4 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCS8275	3.037 ms	0 - 75 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	1.263 ms	0 - 1 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCS8450	1.801 ms	0 - 128 MB	NPU
ResNet101	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.74 ms	0 - 77 MB	NPU
ResNet101	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.8 ms	0 - 88 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCM6690	11.803 ms	0 - 197 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCS9075	1.325 ms	2 - 4 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® SA7255P	3.037 ms	0 - 75 MB	NPU
ResNet101	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	0.799 ms	0 - 74 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® SA8295P	1.953 ms	0 - 73 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCS7790	1.8 ms	0 - 88 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCS8750	0.799 ms	0 - 74 MB	NPU
ResNet101	QNN_DLC	w8a8	Qualcomm® QCS7181	1.258 ms	1 - 1 MB	NPU
ResNet101	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	2.4 ms	0 - 176 MB	NPU
ResNet101	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	5.913 ms	0 - 144 MB	NPU
ResNet101	TFLITE	float	Qualcomm® QCS8275	18.242 ms	0 - 109 MB	NPU
ResNet101	TFLITE	float	Qualcomm® QCS8550 (Proxy)	3.343 ms	0 - 2 MB	NPU
ResNet101	TFLITE	float	Qualcomm® SA8775P	39.242 ms	0 - 20 MB	GPU
ResNet101	TFLITE	float	Qualcomm® SA8650P	39.242 ms	0 - 20 MB	GPU
ResNet101	TFLITE	float	Qualcomm® SA8255P	39.242 ms	0 - 20 MB	GPU
ResNet101	TFLITE	float	Qualcomm® QCS8450	5.913 ms	0 - 144 MB	NPU
ResNet101	TFLITE	float	Snapdragon® 8 Elite Mobile	1.922 ms	0 - 111 MB	NPU
ResNet101	TFLITE	float	Qualcomm® SA8295P	5.573 ms	0 - 77 MB	NPU
ResNet101	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.633 ms	0 - 114 MB	NPU
ResNet101	TFLITE	float	Qualcomm® SA7255P	18.242 ms	0 - 109 MB	NPU
ResNet101	TFLITE	float	Qualcomm® QCS9075	5.287 ms	0 - 88 MB	NPU
ResNet101	TFLITE	float	Qualcomm® QCS8750	1.922 ms	0 - 111 MB	NPU
ResNet101	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.875 ms	0 - 135 MB	NPU
ResNet101	TFLITE	w8a8	Snapdragon® 8 Gen 1 Mobile	1.639 ms	0 - 136 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® QCS6490	4.287 ms	0 - 45 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® QCS8275	2.779 ms	0 - 71 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	1.135 ms	0 - 3 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® SA8775P	39.711 ms	1 - 26 MB	GPU
ResNet101	TFLITE	w8a8	Qualcomm® SA8650P	39.711 ms	1 - 26 MB	GPU
ResNet101	TFLITE	w8a8	Qualcomm® SA8255P	39.711 ms	1 - 26 MB	GPU
ResNet101	TFLITE	w8a8	Qualcomm® QCS8450	1.639 ms	0 - 136 MB	NPU
ResNet101	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.718 ms	0 - 74 MB	NPU
ResNet101	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	1.665 ms	0 - 78 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® QCM6690	12.419 ms	0 - 194 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® QCS9075	1.179 ms	0 - 45 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® SA7255P	2.779 ms	0 - 71 MB	NPU
ResNet101	TFLITE	w8a8	Snapdragon® 8 Elite Mobile	0.735 ms	0 - 71 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® SA8295P	1.809 ms	0 - 68 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® QCS7790	1.665 ms	0 - 78 MB	NPU
ResNet101	TFLITE	w8a8	Qualcomm® QCS8750	0.735 ms	0 - 71 MB	NPU

License

• The license for the original implementation of ResNet101 can be found here.

References

• Deep Residual Learning for Image Recognition
• Source Model Implementation

Community

• Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
• For questions or feedback please reach out to us.