DenseNet-121: Optimized for Qualcomm Devices

Densenet 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 DenseNet-121 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 2.42, ONNX Runtime 1.24.1	Download
ONNX	w8a16	Universal	QAIRT 2.42, ONNX Runtime 1.24.1	Download
QNN_DLC	float	Universal	QAIRT 2.43	Download
QNN_DLC	w8a16	Universal	QAIRT 2.43	Download
QNN_DLC	w8a8	Universal	QAIRT 2.43	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.17.0	Download
TFLITE	w8a8	Universal	QAIRT 2.43, TFLite 2.17.0	Download

For more device-specific assets and performance metrics, visit DenseNet-121 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 DenseNet-121 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: 7.99M
• Model size (float): 30.5 MB
• Model size (w8a16): 8.72 MB
• Model size (w8a8): 8.30 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
DenseNet-121	ONNX	float	Snapdragon® X2 Elite	0.747 ms	16 - 16 MB	NPU
DenseNet-121	ONNX	float	Snapdragon® X Elite	1.734 ms	15 - 15 MB	NPU
DenseNet-121	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.082 ms	0 - 88 MB	NPU
DenseNet-121	ONNX	float	Qualcomm® QCS8550 (Proxy)	1.521 ms	0 - 18 MB	NPU
DenseNet-121	ONNX	float	Qualcomm® QCS9075	2.606 ms	1 - 3 MB	NPU
DenseNet-121	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.784 ms	0 - 46 MB	NPU
DenseNet-121	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.679 ms	0 - 46 MB	NPU
DenseNet-121	ONNX	w8a16	Snapdragon® X2 Elite	1.239 ms	27 - 27 MB	NPU
DenseNet-121	ONNX	w8a16	Snapdragon® X Elite	2.597 ms	25 - 25 MB	NPU
DenseNet-121	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	1.706 ms	0 - 103 MB	NPU
DenseNet-121	ONNX	w8a16	Qualcomm® QCS6490	255.731 ms	32 - 35 MB	CPU
DenseNet-121	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	2.348 ms	0 - 32 MB	NPU
DenseNet-121	ONNX	w8a16	Qualcomm® QCS9075	2.951 ms	0 - 3 MB	NPU
DenseNet-121	ONNX	w8a16	Qualcomm® QCM6690	115.43 ms	20 - 31 MB	CPU
DenseNet-121	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	1.321 ms	0 - 77 MB	NPU
DenseNet-121	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	103.297 ms	30 - 41 MB	CPU
DenseNet-121	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	1.125 ms	0 - 79 MB	NPU
DenseNet-121	QNN_DLC	float	Snapdragon® X2 Elite	0.953 ms	1 - 1 MB	NPU
DenseNet-121	QNN_DLC	float	Snapdragon® X Elite	2.027 ms	1 - 1 MB	NPU
DenseNet-121	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	1.192 ms	0 - 80 MB	NPU
DenseNet-121	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	7.774 ms	1 - 39 MB	NPU
DenseNet-121	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	1.782 ms	1 - 2 MB	NPU
DenseNet-121	QNN_DLC	float	Qualcomm® SA8775P	11.875 ms	1 - 39 MB	NPU
DenseNet-121	QNN_DLC	float	Qualcomm® QCS9075	2.767 ms	1 - 3 MB	NPU
DenseNet-121	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	3.269 ms	0 - 79 MB	NPU
DenseNet-121	QNN_DLC	float	Qualcomm® SA7255P	7.774 ms	1 - 39 MB	NPU
DenseNet-121	QNN_DLC	float	Qualcomm® SA8295P	3.044 ms	0 - 36 MB	NPU
DenseNet-121	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.862 ms	0 - 43 MB	NPU
DenseNet-121	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.707 ms	1 - 43 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Snapdragon® X2 Elite	1.584 ms	0 - 0 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Snapdragon® X Elite	3.458 ms	0 - 0 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	2.146 ms	0 - 90 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Qualcomm® QCS8275 (Proxy)	6.279 ms	0 - 64 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	3.14 ms	0 - 2 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Qualcomm® SA8775P	3.503 ms	0 - 65 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Qualcomm® QCS9075	3.404 ms	0 - 2 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Qualcomm® QCS8450 (Proxy)	5.048 ms	0 - 90 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Qualcomm® SA7255P	6.279 ms	0 - 64 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Qualcomm® SA8295P	5.59 ms	0 - 61 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	1.761 ms	0 - 58 MB	NPU
DenseNet-121	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	1.22 ms	0 - 72 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Snapdragon® X2 Elite	1.15 ms	0 - 0 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Snapdragon® X Elite	2.641 ms	0 - 0 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	1.695 ms	0 - 82 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	4.884 ms	0 - 59 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	2.38 ms	0 - 2 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Qualcomm® SA8775P	2.707 ms	0 - 59 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Qualcomm® QCS9075	2.567 ms	0 - 2 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	2.862 ms	0 - 77 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Qualcomm® SA7255P	4.884 ms	0 - 59 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Qualcomm® SA8295P	3.462 ms	0 - 56 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.046 ms	0 - 58 MB	NPU
DenseNet-121	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.907 ms	0 - 64 MB	NPU
DenseNet-121	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	1.196 ms	0 - 96 MB	NPU
DenseNet-121	TFLITE	float	Qualcomm® QCS8275 (Proxy)	7.803 ms	0 - 55 MB	NPU
DenseNet-121	TFLITE	float	Qualcomm® QCS8550 (Proxy)	1.804 ms	0 - 9 MB	NPU
DenseNet-121	TFLITE	float	Qualcomm® SA8775P	2.726 ms	0 - 57 MB	NPU
DenseNet-121	TFLITE	float	Qualcomm® QCS9075	2.823 ms	0 - 18 MB	NPU
DenseNet-121	TFLITE	float	Qualcomm® QCS8450 (Proxy)	3.293 ms	0 - 92 MB	NPU
DenseNet-121	TFLITE	float	Qualcomm® SA7255P	7.803 ms	0 - 55 MB	NPU
DenseNet-121	TFLITE	float	Qualcomm® SA8295P	3.07 ms	0 - 53 MB	NPU
DenseNet-121	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.85 ms	0 - 58 MB	NPU
DenseNet-121	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.71 ms	0 - 59 MB	NPU
DenseNet-121	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	1.592 ms	0 - 85 MB	NPU
DenseNet-121	TFLITE	w8a8	Qualcomm® QCS6490	87.29 ms	6 - 34 MB	NPU
DenseNet-121	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	4.538 ms	0 - 58 MB	NPU
DenseNet-121	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	2.218 ms	0 - 2 MB	NPU
DenseNet-121	TFLITE	w8a8	Qualcomm® SA8775P	2.602 ms	0 - 62 MB	NPU
DenseNet-121	TFLITE	w8a8	Qualcomm® QCS9075	2.425 ms	0 - 11 MB	NPU
DenseNet-121	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	2.647 ms	0 - 85 MB	NPU
DenseNet-121	TFLITE	w8a8	Qualcomm® SA7255P	4.538 ms	0 - 58 MB	NPU
DenseNet-121	TFLITE	w8a8	Qualcomm® SA8295P	3.313 ms	0 - 57 MB	NPU
DenseNet-121	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.009 ms	0 - 57 MB	NPU
DenseNet-121	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	26.363 ms	6 - 172 MB	NPU
DenseNet-121	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.862 ms	0 - 65 MB	NPU

License

• The license for the original implementation of DenseNet-121 can be found here.

References

• Densely Connected Convolutional Networks
• 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.