Depth-Anything-V2: Optimized for Qualcomm Devices

Depth Anything is designed for estimating depth at each point in an image.

This is based on the implementation of Depth-Anything-V2 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 2.50, ONNX Runtime 1.30.0	Download
ONNX	w8a16	Universal	QAIRT 2.50, ONNX Runtime 1.30.0	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
QNN_DLC	w8a16	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	Download

For more device-specific assets and performance metrics, visit Depth-Anything-V2 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 Depth-Anything-V2 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.depth_estimation

Model Stats:

• Input resolution: 518x518
• Model checkpoint: DepthAnything_V2_Small
• Model size (float): 94.3 MB
• Number of parameters: 24.7M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Depth-Anything-V2	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	23.061 ms	2 - 385 MB	NPU
Depth-Anything-V2	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	25.799 ms	2 - 375 MB	NPU
Depth-Anything-V2	ONNX	float	Snapdragon® X2 Elite	20.348 ms	5 - 5 MB	NPU
Depth-Anything-V2	ONNX	float	Snapdragon® X Elite	47.567 ms	48 - 48 MB	NPU
Depth-Anything-V2	ONNX	float	Snapdragon® 8 Gen 3 Mobile	34.164 ms	0 - 490 MB	NPU
Depth-Anything-V2	ONNX	float	Snapdragon® 8 Gen 1 Mobile	71.72 ms	0 - 502 MB	NPU
Depth-Anything-V2	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	52.848 ms	3 - 10 MB	NPU
Depth-Anything-V2	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	46.692 ms	4 - 7 MB	NPU
Depth-Anything-V2	ONNX	float	Qualcomm® QCS8450	71.72 ms	0 - 502 MB	NPU
Depth-Anything-V2	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	52.242 ms	3 - 9 MB	NPU
Depth-Anything-V2	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	47.567 ms	48 - 48 MB	NPU
Depth-Anything-V2	ONNX	float	Qualcomm® Dragonwing™ Q-8750	25.799 ms	2 - 375 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	11.919 ms	1 - 421 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	14.42 ms	2 - 412 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Snapdragon® X2 Elite	11.381 ms	4 - 4 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Snapdragon® X Elite	26.798 ms	27 - 27 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	19.293 ms	3 - 512 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	33.528 ms	3 - 515 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	85.561 ms	2 - 5 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-8275	25.144 ms	2 - 6 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	26.726 ms	0 - 34 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® QCS8450	33.528 ms	3 - 515 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	27.322 ms	2 - 5 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	26.798 ms	27 - 27 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	211.465 ms	0 - 532 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	34.111 ms	3 - 462 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	14.42 ms	2 - 412 MB	NPU
Depth-Anything-V2	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	34.111 ms	3 - 462 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	16.732 ms	3 - 371 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	21.464 ms	0 - 355 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Snapdragon® X2 Elite	20.743 ms	3 - 3 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Snapdragon® X Elite	45.041 ms	3 - 3 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	30.772 ms	3 - 467 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	66.405 ms	3 - 467 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	48.071 ms	3 - 10 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	43.488 ms	3 - 159 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Qualcomm® QCS8450	66.405 ms	3 - 467 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	48.545 ms	3 - 9 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	45.041 ms	3 - 3 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	21.464 ms	0 - 355 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Qualcomm® SA7255P	124.242 ms	0 - 352 MB	NPU
Depth-Anything-V2	QNN_DLC	float	Qualcomm® SA8295P	69.92 ms	0 - 352 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	13.391 ms	2 - 402 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	15.138 ms	2 - 392 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Snapdragon® X2 Elite	13.807 ms	2 - 2 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Snapdragon® X Elite	31.196 ms	2 - 2 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	22.348 ms	0 - 479 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	29.981 ms	2 - 7 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	29.886 ms	2 - 4 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® SA8650P	140.741 ms	2 - 2002 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® SA8255P	140.741 ms	2 - 2002 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	31.271 ms	1 - 5 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	31.196 ms	2 - 2 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-6690	300.713 ms	2 - 504 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-7790	46.609 ms	2 - 440 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	15.138 ms	2 - 392 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Qualcomm® SA7255P	62.347 ms	2 - 389 MB	NPU
Depth-Anything-V2	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	46.609 ms	2 - 440 MB	NPU
Depth-Anything-V2	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	16.799 ms	0 - 378 MB	NPU
Depth-Anything-V2	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	21.505 ms	0 - 365 MB	NPU
Depth-Anything-V2	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	30.652 ms	0 - 469 MB	NPU
Depth-Anything-V2	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	67.65 ms	1 - 468 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	48.1 ms	1 - 58 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	42.815 ms	1 - 9 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® SA8775P	48.359 ms	1 - 361 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® SA8650P	48.359 ms	1 - 361 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® SA8255P	48.359 ms	1 - 361 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® QCS8450	67.65 ms	1 - 468 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	48.713 ms	1 - 58 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	21.505 ms	0 - 365 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® SA7255P	124.349 ms	1 - 361 MB	NPU
Depth-Anything-V2	TFLITE	float	Qualcomm® SA8295P	69.934 ms	1 - 358 MB	NPU

License

• The license for the original implementation of Depth-Anything-V2 can be found here.

References

• Depth Anything V2
• 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.