Midas-V2: Optimized for Mobile Deployment
Deep Convolutional Neural Network model for depth estimation
Midas is designed for estimating depth at each point in an image.
This model is an implementation of Midas-V2 found here.
This repository provides scripts to run Midas-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: MiDaS_small
- Input resolution: 256x256
- Number of parameters: 16.6M
- Model size (float): 63.2 MB
- Model size (w8a8): 16.9 MB
| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model |
|---|---|---|---|---|---|---|---|---|
| Midas-V2 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 13.185 ms | 0 - 45 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 11.965 ms | 1 - 30 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 4.94 ms | 0 - 60 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 7.452 ms | 0 - 39 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 3.28 ms | 0 - 316 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 3.0 ms | 1 - 11 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 3.058 ms | 0 - 128 MB | NPU | Midas-V2.onnx.zip |
| Midas-V2 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 4.668 ms | 0 - 45 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 4.179 ms | 41 - 70 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 13.185 ms | 0 - 45 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 11.965 ms | 1 - 30 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 3.304 ms | 0 - 318 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 3.008 ms | 0 - 15 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 5.834 ms | 1 - 33 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 5.33 ms | 1 - 32 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 3.296 ms | 0 - 309 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 3.007 ms | 0 - 13 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 4.668 ms | 0 - 45 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 4.179 ms | 41 - 70 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 2.318 ms | 0 - 70 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 2.081 ms | 1 - 44 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.107 ms | 0 - 43 MB | NPU | Midas-V2.onnx.zip |
| Midas-V2 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 1.822 ms | 0 - 51 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 1.535 ms | 1 - 38 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 1.684 ms | 0 - 33 MB | NPU | Midas-V2.onnx.zip |
| Midas-V2 | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 1.444 ms | 0 - 49 MB | NPU | Midas-V2.tflite |
| Midas-V2 | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 1.323 ms | 0 - 36 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 1.428 ms | 1 - 32 MB | NPU | Midas-V2.onnx.zip |
| Midas-V2 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 3.209 ms | 193 - 193 MB | NPU | Midas-V2.dlc |
| Midas-V2 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.906 ms | 36 - 36 MB | NPU | Midas-V2.onnx.zip |
| Midas-V2 | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | TFLITE | 3.671 ms | 0 - 28 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | QNN_DLC | 4.056 ms | 0 - 102 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 2.584 ms | 0 - 33 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 2.893 ms | 0 - 34 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.427 ms | 0 - 52 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.838 ms | 0 - 52 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.073 ms | 0 - 147 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.28 ms | 0 - 134 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.416 ms | 0 - 34 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.599 ms | 0 - 33 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 16.06 ms | 0 - 3 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 2.584 ms | 0 - 33 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 2.893 ms | 0 - 34 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 1.083 ms | 0 - 145 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.271 ms | 0 - 135 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 1.954 ms | 0 - 38 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 2.205 ms | 0 - 40 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 1.077 ms | 0 - 140 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.285 ms | 0 - 140 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 1.416 ms | 0 - 34 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.599 ms | 0 - 33 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.763 ms | 0 - 64 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.917 ms | 0 - 64 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 0.596 ms | 0 - 47 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.676 ms | 0 - 42 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | TFLITE | 1.486 ms | 0 - 50 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 1.578 ms | 0 - 50 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 0.526 ms | 0 - 40 MB | NPU | Midas-V2.tflite |
| Midas-V2 | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 0.56 ms | 0 - 42 MB | NPU | Midas-V2.dlc |
| Midas-V2 | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.44 ms | 143 - 143 MB | NPU | Midas-V2.dlc |
Installation
Install the package via pip:
# NOTE: 3.10 <= PYTHON_VERSION < 3.14 is supported.
pip install "qai-hub-models[midas]"
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.midas.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.midas.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.midas.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.midas 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.midas.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.midas.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 Midas-V2's performance across various devices here. Explore all available models on Qualcomm® AI Hub
License
- The license for the original implementation of Midas-V2 can be found here.
- The license for the compiled assets for on-device deployment 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.
- Downloads last month
- 1,445