Depth-Anything: 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 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 1.24.3 | Download |
| ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.3 | Download |
| QNN_DLC | float | Universal | QAIRT 2.45 | Download |
| QNN_DLC | w8a16 | Universal | QAIRT 2.45 | Download |
| TFLITE | float | Universal | QAIRT 2.45 | Download |
For more device-specific assets and performance metrics, visit Depth-Anything 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 on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.depth_estimation
Model Stats:
- Model checkpoint: DepthAnything_Small
- Input resolution: 518x518
- Number of parameters: 24.7M
- Model size (float): 94.3 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| Depth-Anything | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 18.383 ms | 1 - 416 MB | NPU |
| Depth-Anything | ONNX | float | Snapdragon® 8 Elite Mobile | 22.849 ms | 3 - 405 MB | NPU |
| Depth-Anything | ONNX | float | Snapdragon® X2 Elite | 18.672 ms | 48 - 48 MB | NPU |
| Depth-Anything | ONNX | float | Snapdragon® X Elite | 45.086 ms | 47 - 47 MB | NPU |
| Depth-Anything | ONNX | float | Snapdragon® X Elite | 45.086 ms | 47 - 47 MB | NPU |
| Depth-Anything | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 31.93 ms | 0 - 583 MB | NPU |
| Depth-Anything | ONNX | float | Qualcomm® QCS8550 (Proxy) | 43.832 ms | 0 - 75 MB | NPU |
| Depth-Anything | ONNX | float | Qualcomm® QCS9075 | 54.875 ms | 3 - 9 MB | NPU |
| Depth-Anything | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 22.849 ms | 3 - 405 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 23.198 ms | 3 - 489 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 26.251 ms | 3 - 482 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® X2 Elite | 24.136 ms | 28 - 28 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® X Elite | 44.062 ms | 27 - 27 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® X Elite | 44.062 ms | 27 - 27 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 29.416 ms | 3 - 770 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Qualcomm® QCS6490 | 3440.152 ms | 204 - 209 MB | CPU |
| Depth-Anything | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 42.267 ms | 0 - 34 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Qualcomm® QCS9075 | 66.209 ms | 2 - 5 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Qualcomm® QCM6690 | 1777.961 ms | 200 - 215 MB | CPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 26.251 ms | 3 - 482 MB | NPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 1698.961 ms | 201 - 213 MB | CPU |
| Depth-Anything | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 1698.961 ms | 201 - 213 MB | CPU |
| Depth-Anything | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 19.769 ms | 3 - 381 MB | NPU |
| Depth-Anything | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 24.876 ms | 0 - 384 MB | NPU |
| Depth-Anything | QNN_DLC | float | Snapdragon® X2 Elite | 20.241 ms | 3 - 3 MB | NPU |
| Depth-Anything | QNN_DLC | float | Snapdragon® X Elite | 51.188 ms | 3 - 3 MB | NPU |
| Depth-Anything | QNN_DLC | float | Snapdragon® X Elite | 51.188 ms | 3 - 3 MB | NPU |
| Depth-Anything | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 38.673 ms | 3 - 513 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 142.191 ms | 0 - 387 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 51.449 ms | 3 - 5 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® SA8775P | 56.617 ms | 1 - 371 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® SA8775P | 56.617 ms | 1 - 371 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® SA8775P | 56.617 ms | 1 - 371 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® QCS9075 | 62.288 ms | 3 - 9 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 92.355 ms | 3 - 489 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® SA7255P | 142.191 ms | 0 - 387 MB | NPU |
| Depth-Anything | QNN_DLC | float | Qualcomm® SA8295P | 82.563 ms | 0 - 362 MB | NPU |
| Depth-Anything | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 24.876 ms | 0 - 384 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 81.92 ms | 2 - 2004 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 93.997 ms | 2 - 2009 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 70.787 ms | 2 - 2 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Snapdragon® X Elite | 138.296 ms | 2 - 2 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Snapdragon® X Elite | 138.296 ms | 2 - 2 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 115.18 ms | 0 - 2298 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 266.457 ms | 2 - 1765 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 141.946 ms | 2 - 460 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Qualcomm® SA8775P | 141.455 ms | 2 - 1999 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Qualcomm® SA8775P | 141.455 ms | 2 - 1999 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Qualcomm® SA8775P | 141.455 ms | 2 - 1999 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 141.614 ms | 2 - 5 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Qualcomm® SA7255P | 266.457 ms | 2 - 1765 MB | NPU |
| Depth-Anything | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 93.997 ms | 2 - 2009 MB | NPU |
| Depth-Anything | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 19.794 ms | 1 - 389 MB | NPU |
| Depth-Anything | TFLITE | float | Snapdragon® 8 Elite Mobile | 24.888 ms | 0 - 395 MB | NPU |
| Depth-Anything | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 38.9 ms | 1 - 521 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 142.497 ms | 1 - 395 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 50.891 ms | 1 - 7 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® SA8775P | 57.067 ms | 1 - 394 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® SA8775P | 57.067 ms | 1 - 394 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® SA8775P | 57.067 ms | 1 - 394 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® QCS9075 | 61.52 ms | 1 - 57 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 92.657 ms | 0 - 487 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® SA7255P | 142.497 ms | 1 - 395 MB | NPU |
| Depth-Anything | TFLITE | float | Qualcomm® SA8295P | 82.51 ms | 1 - 369 MB | NPU |
| Depth-Anything | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 24.888 ms | 0 - 395 MB | NPU |
License
- The license for the original implementation of Depth-Anything can be found here.
References
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.
