Depth-Anything-V2: Optimized for Mobile Deployment
Deep Convolutional Neural Network model for depth estimation
Depth Anything is designed for estimating depth at each point in an image.
This model is an implementation of Depth-Anything-V2 found here.
This repository provides scripts to run Depth-Anything-V2 on Qualcomm® devices. More details on model performance across various devices, can be found here.
Model Details
- Model Type: Model_use_case.depth_estimation
- Model Stats:
- Model checkpoint: DepthAnything_V2_Small
- Input resolution: 518x518
- Number of parameters: 24.7M
- Model size (float): 94.3 MB
| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model |
|---|---|---|---|---|---|---|---|---|
| Depth-Anything-V2 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 197.472 ms | 0 - 170 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 179.667 ms | 0 - 185 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 119.407 ms | 1 - 168 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 121.189 ms | 3 - 183 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 85.22 ms | 1 - 38 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 76.512 ms | 3 - 47 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 77.939 ms | 4 - 31 MB | NPU | Depth-Anything-V2.onnx.zip |
| Depth-Anything-V2 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 92.215 ms | 0 - 170 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 81.904 ms | 0 - 185 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 197.472 ms | 0 - 170 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 179.667 ms | 0 - 185 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 84.912 ms | 1 - 38 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 76.562 ms | 4 - 52 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 132.909 ms | 1 - 165 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 115.862 ms | 0 - 181 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 85.857 ms | 0 - 36 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 76.646 ms | 3 - 48 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 92.215 ms | 0 - 170 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 81.904 ms | 0 - 185 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 59.518 ms | 1 - 175 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 56.746 ms | 3 - 193 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 59.939 ms | 0 - 415 MB | NPU | Depth-Anything-V2.onnx.zip |
| Depth-Anything-V2 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 47.524 ms | 0 - 170 MB | NPU | Depth-Anything-V2.tflite |
| Depth-Anything-V2 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 40.587 ms | 3 - 186 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 45.998 ms | 4 - 511 MB | NPU | Depth-Anything-V2.onnx.zip |
| Depth-Anything-V2 | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 35.162 ms | 3 - 204 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 41.624 ms | 0 - 456 MB | NPU | Depth-Anything-V2.onnx.zip |
| Depth-Anything-V2 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 76.854 ms | 87 - 87 MB | NPU | Depth-Anything-V2.dlc |
| Depth-Anything-V2 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 77.348 ms | 62 - 62 MB | NPU | Depth-Anything-V2.onnx.zip |
Installation
Install the package via pip:
# NOTE: 3.10 <= PYTHON_VERSION < 3.14 is supported.
pip install "qai-hub-models[depth-anything-v2]"
Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
Sign-in to Qualcomm® AI Hub Workbench with your
Qualcomm® ID. Once signed in navigate to Account -> Settings -> API Token.
With this API token, you can configure your client to run models on the cloud hosted devices.
qai-hub configure --api_token API_TOKEN
Navigate to docs for more information.
Demo off target
The package contains a simple end-to-end demo that downloads pre-trained weights and runs this model on a sample input.
python -m qai_hub_models.models.depth_anything_v2.demo
The above demo runs a reference implementation of pre-processing, model inference, and post processing.
NOTE: If you want running in a Jupyter Notebook or Google Colab like environment, please add the following to your cell (instead of the above).
%run -m qai_hub_models.models.depth_anything_v2.demo
Run model on a cloud-hosted device
In addition to the demo, you can also run the model on a cloud-hosted Qualcomm® device. This script does the following:
- Performance check on-device on a cloud-hosted device
- Downloads compiled assets that can be deployed on-device for Android.
- Accuracy check between PyTorch and on-device outputs.
python -m qai_hub_models.models.depth_anything_v2.export
How does this work?
This export script leverages Qualcomm® AI Hub to optimize, validate, and deploy this model on-device. Lets go through each step below in detail:
Step 1: Compile model for on-device deployment
To compile a PyTorch model for on-device deployment, we first trace the model
in memory using the jit.trace and then call the submit_compile_job API.
import torch
import qai_hub as hub
from qai_hub_models.models.depth_anything_v2 import Model
# Load the model
torch_model = Model.from_pretrained()
# Device
device = hub.Device("Samsung Galaxy S25")
# Trace model
input_shape = torch_model.get_input_spec()
sample_inputs = torch_model.sample_inputs()
pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
# Compile model on a specific device
compile_job = hub.submit_compile_job(
model=pt_model,
device=device,
input_specs=torch_model.get_input_spec(),
)
# Get target model to run on-device
target_model = compile_job.get_target_model()
Step 2: Performance profiling on cloud-hosted device
After compiling models from step 1. Models can be profiled model on-device using the
target_model. Note that this scripts runs the model on a device automatically
provisioned in the cloud. Once the job is submitted, you can navigate to a
provided job URL to view a variety of on-device performance metrics.
profile_job = hub.submit_profile_job(
model=target_model,
device=device,
)
Step 3: Verify on-device accuracy
To verify the accuracy of the model on-device, you can run on-device inference on sample input data on the same cloud hosted device.
input_data = torch_model.sample_inputs()
inference_job = hub.submit_inference_job(
model=target_model,
device=device,
inputs=input_data,
)
on_device_output = inference_job.download_output_data()
With the output of the model, you can compute like PSNR, relative errors or spot check the output with expected output.
Note: This on-device profiling and inference requires access to Qualcomm® AI Hub Workbench. Sign up for access.
Run demo on a cloud-hosted device
You can also run the demo on-device.
python -m qai_hub_models.models.depth_anything_v2.demo --eval-mode on-device
NOTE: If you want running in a Jupyter Notebook or Google Colab like environment, please add the following to your cell (instead of the above).
%run -m qai_hub_models.models.depth_anything_v2.demo -- --eval-mode on-device
Deploying compiled model to Android
The models can be deployed using multiple runtimes:
TensorFlow Lite (
.tfliteexport): This tutorial provides a guide to deploy the .tflite model in an Android application.QNN (
.soexport ): This sample app provides instructions on how to use the.soshared library in an Android application.
View on Qualcomm® AI Hub
Get more details on Depth-Anything-V2's performance across various devices here. Explore all available models on Qualcomm® AI Hub
License
- The license for the original implementation of Depth-Anything-V2 can be found here.
- The license for the compiled assets for on-device deployment 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.
- Downloads last month
- 1,335