Yolo-v7: Optimized for Mobile Deployment
Real-time object detection optimized for mobile and edge
YoloV7 is a machine learning model that predicts bounding boxes and classes of objects in an image.
This model is an implementation of Yolo-v7 found here.
This repository provides scripts to run Yolo-v7 on Qualcomm® devices. More details on model performance across various devices, can be found here.
WARNING: The model assets are not readily available for download due to licensing restrictions.
Model Details
- Model Type: Model_use_case.object_detection
- Model Stats:
- Model checkpoint: YoloV7 Tiny
- Input resolution: 640x640
- Number of parameters: 6.24M
- Model size (float): 23.8 MB
- Model size (w8a8): 6.23 MB
- Model size (w8a16): 6.66 MB
| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model |
|---|---|---|---|---|---|---|---|---|
| Yolo-v7 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 26.953 ms | 1 - 127 MB | NPU | -- |
| Yolo-v7 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 21.079 ms | 2 - 131 MB | NPU | -- |
| Yolo-v7 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 11.046 ms | 1 - 49 MB | NPU | -- |
| Yolo-v7 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 13.655 ms | 5 - 45 MB | NPU | -- |
| Yolo-v7 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 10.855 ms | 0 - 112 MB | NPU | -- |
| Yolo-v7 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 10.48 ms | 5 - 27 MB | NPU | -- |
| Yolo-v7 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 10.107 ms | 0 - 133 MB | NPU | -- |
| Yolo-v7 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 12.508 ms | 1 - 127 MB | NPU | -- |
| Yolo-v7 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 10.433 ms | 3 - 130 MB | NPU | -- |
| Yolo-v7 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 26.953 ms | 1 - 127 MB | NPU | -- |
| Yolo-v7 | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 21.079 ms | 2 - 131 MB | NPU | -- |
| Yolo-v7 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 10.895 ms | 0 - 127 MB | NPU | -- |
| Yolo-v7 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 10.517 ms | 5 - 21 MB | NPU | -- |
| Yolo-v7 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 12.75 ms | 1 - 40 MB | NPU | -- |
| Yolo-v7 | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 11.897 ms | 2 - 44 MB | NPU | -- |
| Yolo-v7 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 10.84 ms | 0 - 135 MB | NPU | -- |
| Yolo-v7 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 10.473 ms | 5 - 24 MB | NPU | -- |
| Yolo-v7 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 12.508 ms | 1 - 127 MB | NPU | -- |
| Yolo-v7 | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 10.433 ms | 3 - 130 MB | NPU | -- |
| Yolo-v7 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 7.496 ms | 0 - 254 MB | NPU | -- |
| Yolo-v7 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 7.063 ms | 4 - 298 MB | NPU | -- |
| Yolo-v7 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 6.94 ms | 5 - 321 MB | NPU | -- |
| Yolo-v7 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 6.251 ms | 1 - 109 MB | NPU | -- |
| Yolo-v7 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 4.977 ms | 5 - 133 MB | NPU | -- |
| Yolo-v7 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 6.938 ms | 0 - 125 MB | NPU | -- |
| Yolo-v7 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 11.128 ms | 222 - 222 MB | NPU | -- |
| Yolo-v7 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 10.752 ms | 9 - 9 MB | NPU | -- |
| Yolo-v7 | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 14.805 ms | 2 - 38 MB | NPU | -- |
| Yolo-v7 | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 12.172 ms | 2 - 57 MB | NPU | -- |
| Yolo-v7 | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 9.574 ms | 2 - 14 MB | NPU | -- |
| Yolo-v7 | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 113.132 ms | 29 - 378 MB | NPU | -- |
| Yolo-v7 | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 10.16 ms | 2 - 39 MB | NPU | -- |
| Yolo-v7 | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 23.692 ms | 2 - 44 MB | NPU | -- |
| Yolo-v7 | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 227.488 ms | 60 - 73 MB | CPU | -- |
| Yolo-v7 | w8a16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 180.52 ms | 79 - 91 MB | CPU | -- |
| Yolo-v7 | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 14.805 ms | 2 - 38 MB | NPU | -- |
| Yolo-v7 | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 9.583 ms | 2 - 14 MB | NPU | -- |
| Yolo-v7 | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 12.001 ms | 2 - 51 MB | NPU | -- |
| Yolo-v7 | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 9.577 ms | 2 - 14 MB | NPU | -- |
| Yolo-v7 | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 10.16 ms | 2 - 39 MB | NPU | -- |
| Yolo-v7 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 5.166 ms | 2 - 51 MB | NPU | -- |
| Yolo-v7 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 93.143 ms | 39 - 1739 MB | NPU | -- |
| Yolo-v7 | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 5.559 ms | 2 - 46 MB | NPU | -- |
| Yolo-v7 | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 82.7 ms | 51 - 1813 MB | NPU | -- |
| Yolo-v7 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 9.957 ms | 17 - 17 MB | NPU | -- |
| Yolo-v7 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 109.781 ms | 56 - 56 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 5.574 ms | 0 - 28 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 4.288 ms | 1 - 30 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 3.421 ms | 0 - 48 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 2.799 ms | 1 - 46 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 2.67 ms | 0 - 32 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.915 ms | 1 - 13 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 3.608 ms | 0 - 48 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 3.077 ms | 0 - 28 MB | NPU | -- |
| Yolo-v7 | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 2.375 ms | 0 - 30 MB | NPU | -- |
| Yolo-v7 | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 20.561 ms | 8 - 54 MB | NPU | -- |
| Yolo-v7 | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 8.568 ms | 1 - 39 MB | NPU | -- |
| Yolo-v7 | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 57.194 ms | 36 - 51 MB | CPU | -- |
| Yolo-v7 | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 127.355 ms | 17 - 47 MB | GPU | -- |
| Yolo-v7 | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 51.618 ms | 35 - 45 MB | CPU | -- |
| Yolo-v7 | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 5.574 ms | 0 - 28 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 4.288 ms | 1 - 30 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 2.671 ms | 0 - 33 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.917 ms | 1 - 17 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 4.322 ms | 0 - 35 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 3.281 ms | 1 - 37 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 2.625 ms | 5 - 36 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.917 ms | 2 - 15 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 3.077 ms | 0 - 28 MB | NPU | -- |
| Yolo-v7 | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 2.375 ms | 0 - 30 MB | NPU | -- |
| Yolo-v7 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.748 ms | 0 - 43 MB | NPU | -- |
| Yolo-v7 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.293 ms | 1 - 48 MB | NPU | -- |
| Yolo-v7 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.523 ms | 0 - 208 MB | NPU | -- |
| Yolo-v7 | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 1.343 ms | 0 - 31 MB | NPU | -- |
| Yolo-v7 | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.981 ms | 1 - 34 MB | NPU | -- |
| Yolo-v7 | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 2.122 ms | 0 - 148 MB | NPU | -- |
| Yolo-v7 | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 2.163 ms | 22 - 22 MB | NPU | -- |
| Yolo-v7 | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.648 ms | 5 - 5 MB | NPU | -- |
Installation
Install the package via pip:
pip install "qai-hub-models[yolov7]"
Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
Sign-in to Qualcomm® AI Hub 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.yolov7.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.yolov7.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.yolov7.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.yolov7 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. 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.yolov7.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.yolov7.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 Yolo-v7's performance across various devices here. Explore all available models on Qualcomm® AI Hub
License
- The license for the original implementation of Yolo-v7 can be found here.
- The license for the compiled assets for on-device deployment can be found here
References
- YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors
- 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.