Midas-V2: Optimized for Qualcomm Devices

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

This is based on the implementation of Midas-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 2.45, ONNX Runtime 1.27.1	Download
ONNX	w8a8	Universal	QAIRT 2.45, ONNX Runtime 1.27.1	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 Midas-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 Midas-V2 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.depth_estimation

Model Stats:

• Input resolution: 256x256
• Model checkpoint: MiDaS_small
• Model size (float): 63.2 MB
• Model size (w8a8): 16.9 MB
• Number of parameters: 16.6M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Midas-V2	ONNX	float	Snapdragon® X2 Elite	1.317 ms	2 - 2 MB	NPU
Midas-V2	ONNX	float	Snapdragon® X Elite	2.763 ms	36 - 36 MB	NPU
Midas-V2	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.933 ms	0 - 196 MB	NPU
Midas-V2	ONNX	float	Snapdragon® 8 Gen 1 Mobile	6.51 ms	1 - 199 MB	NPU
Midas-V2	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	3.925 ms	1 - 5 MB	NPU
Midas-V2	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	2.727 ms	0 - 41 MB	NPU
Midas-V2	ONNX	float	Qualcomm® QCS8450	6.51 ms	1 - 199 MB	NPU
Midas-V2	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	3.664 ms	1 - 4 MB	NPU
Midas-V2	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	2.763 ms	36 - 36 MB	NPU
Midas-V2	ONNX	float	Qualcomm® Dragonwing™ Q-8750	1.47 ms	0 - 163 MB	NPU
Midas-V2	ONNX	float	Snapdragon® 8 Elite Mobile	1.47 ms	0 - 163 MB	NPU
Midas-V2	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	1.26 ms	0 - 163 MB	NPU
Midas-V2	ONNX	w8a8	Snapdragon® X2 Elite	0.616 ms	1 - 1 MB	NPU
Midas-V2	ONNX	w8a8	Snapdragon® X Elite	1.333 ms	19 - 19 MB	NPU
Midas-V2	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.934 ms	0 - 102 MB	NPU
Midas-V2	ONNX	w8a8	Snapdragon® 8 Gen 1 Mobile	1.857 ms	0 - 94 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® Dragonwing™ QCS6490	4.669 ms	0 - 3 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.378 ms	0 - 4 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.328 ms	0 - 22 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® QCS8450	1.857 ms	0 - 94 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.482 ms	0 - 3 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-X7181	1.333 ms	19 - 19 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-6690	8.898 ms	0 - 187 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-7790	1.647 ms	0 - 181 MB	NPU
Midas-V2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-8750	0.762 ms	0 - 61 MB	NPU
Midas-V2	ONNX	w8a8	Snapdragon® 8 Elite Mobile	0.762 ms	0 - 61 MB	NPU
Midas-V2	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.661 ms	0 - 60 MB	NPU
Midas-V2	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	1.647 ms	0 - 181 MB	NPU
Midas-V2	QNN_DLC	float	Snapdragon® X2 Elite	1.543 ms	1 - 1 MB	NPU
Midas-V2	QNN_DLC	float	Snapdragon® X Elite	3.271 ms	1 - 1 MB	NPU
Midas-V2	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	2.114 ms	0 - 193 MB	NPU
Midas-V2	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	6.831 ms	0 - 192 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	4.091 ms	1 - 4 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	3.044 ms	1 - 3 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® SA8775P	4.261 ms	1 - 160 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® SA8650P	4.261 ms	1 - 160 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® SA8255P	4.261 ms	1 - 160 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® QCS8450	6.831 ms	0 - 192 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	4.017 ms	1 - 3 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	3.271 ms	1 - 1 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	1.603 ms	0 - 160 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® SA7255P	12.063 ms	1 - 158 MB	NPU
Midas-V2	QNN_DLC	float	Qualcomm® SA8295P	5.45 ms	1 - 155 MB	NPU
Midas-V2	QNN_DLC	float	Snapdragon® 8 Elite Mobile	1.603 ms	0 - 160 MB	NPU
Midas-V2	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.297 ms	1 - 164 MB	NPU
Midas-V2	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.714 ms	0 - 0 MB	NPU
Midas-V2	QNN_DLC	w8a8	Snapdragon® X Elite	1.451 ms	0 - 0 MB	NPU
Midas-V2	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.926 ms	0 - 82 MB	NPU
Midas-V2	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	1.91 ms	0 - 84 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS6490	5.65 ms	0 - 2 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.337 ms	0 - 3 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.32 ms	0 - 2 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® SA8775P	1.585 ms	0 - 54 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® SA8650P	1.585 ms	0 - 54 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® SA8255P	1.585 ms	0 - 54 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® QCS8450	1.91 ms	0 - 84 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.439 ms	0 - 2 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-X7181	1.451 ms	0 - 0 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-6690	9.172 ms	0 - 179 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-7790	1.616 ms	0 - 172 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-8750	0.694 ms	0 - 47 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® SA7255P	2.88 ms	0 - 48 MB	NPU
Midas-V2	QNN_DLC	w8a8	Qualcomm® SA8295P	2.237 ms	0 - 47 MB	NPU
Midas-V2	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	0.694 ms	0 - 47 MB	NPU
Midas-V2	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.578 ms	0 - 53 MB	NPU
Midas-V2	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.616 ms	0 - 172 MB	NPU
Midas-V2	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	2.116 ms	0 - 227 MB	NPU
Midas-V2	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	6.806 ms	0 - 215 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	4.11 ms	0 - 39 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	3.018 ms	0 - 3 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® SA8775P	4.301 ms	0 - 174 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® SA8650P	4.301 ms	0 - 174 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® SA8255P	4.301 ms	0 - 174 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® QCS8450	6.806 ms	0 - 215 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	4.027 ms	0 - 39 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	1.607 ms	0 - 174 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® SA7255P	12.109 ms	0 - 171 MB	NPU
Midas-V2	TFLITE	float	Qualcomm® SA8295P	5.461 ms	0 - 156 MB	NPU
Midas-V2	TFLITE	float	Snapdragon® 8 Elite Mobile	1.607 ms	0 - 174 MB	NPU
Midas-V2	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.295 ms	0 - 174 MB	NPU
Midas-V2	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.755 ms	0 - 79 MB	NPU
Midas-V2	TFLITE	w8a8	Snapdragon® 8 Gen 1 Mobile	1.619 ms	0 - 84 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS6490	5.187 ms	0 - 29 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.146 ms	0 - 21 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.085 ms	0 - 19 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® SA8775P	1.38 ms	0 - 50 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® SA8650P	1.38 ms	0 - 50 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® SA8255P	1.38 ms	0 - 50 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® QCS8450	1.619 ms	0 - 84 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.218 ms	0 - 20 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-6690	8.385 ms	0 - 179 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-7790	1.383 ms	0 - 169 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-8750	0.606 ms	0 - 50 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® SA7255P	2.513 ms	0 - 46 MB	NPU
Midas-V2	TFLITE	w8a8	Qualcomm® SA8295P	1.992 ms	0 - 44 MB	NPU
Midas-V2	TFLITE	w8a8	Snapdragon® 8 Elite Mobile	0.606 ms	0 - 50 MB	NPU
Midas-V2	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.508 ms	0 - 47 MB	NPU
Midas-V2	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	1.383 ms	0 - 169 MB	NPU

License

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

References

• Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer
• 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.