--- library_name: pytorch license: gpl-3.0 tags: - real_time - quantized - android pipeline_tag: object-detection --- ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/yolov7_quantized/web-assets/model_demo.png) # Yolo-v7-Quantized: Optimized for Mobile Deployment ## Quantized 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 post-training quantized to int8 using samples from the COCO dataset. This model is an implementation of Yolo-v7-Quantized found [here](https://github.com/WongKinYiu/yolov7/). More details on model performance across various devices, can be found [here](https://aihub.qualcomm.com/models/yolov7_quantized). ### Model Details - **Model Type:** Object detection - **Model Stats:** - Model checkpoint: YoloV7 Tiny - Input resolution: 640x640 - Number of parameters: 6.24M - Model size: 6.23 MB - Precision: w8a8 (8-bit weights, 8-bit activations) | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model |---|---|---|---|---|---|---|---|---| | Yolo-v7-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 2.503 ms | 0 - 28 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 2.75 ms | 1 - 4 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 5.219 ms | 0 - 24 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 1.695 ms | 0 - 44 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 1.827 ms | 1 - 18 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 3.758 ms | 1 - 68 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.658 ms | 0 - 25 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 2.183 ms | 1 - 26 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 3.583 ms | 0 - 56 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA7255P ADP | SA7255P | TFLITE | 16.484 ms | 0 - 21 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA7255P ADP | SA7255P | QNN | 16.192 ms | 1 - 11 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 2.498 ms | 0 - 29 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA8255 (Proxy) | SA8255P Proxy | QNN | 2.75 ms | 1 - 11 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA8295P ADP | SA8295P | TFLITE | 3.823 ms | 0 - 26 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA8295P ADP | SA8295P | QNN | 4.324 ms | 1 - 19 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 2.501 ms | 0 - 29 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA8650 (Proxy) | SA8650P Proxy | QNN | 2.739 ms | 1 - 3 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA8775P ADP | SA8775P | TFLITE | 3.552 ms | 0 - 21 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | SA8775P ADP | SA8775P | QNN | 3.898 ms | 1 - 11 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | TFLITE | 10.192 ms | 0 - 33 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | QNN | 7.898 ms | 1 - 15 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | RB5 (Proxy) | QCS8250 Proxy | TFLITE | 59.37 ms | 15 - 53 MB | INT8 | GPU | -- | | Yolo-v7-Quantized | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 16.484 ms | 0 - 21 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 16.192 ms | 1 - 11 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 2.492 ms | 0 - 29 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 2.745 ms | 1 - 4 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 3.552 ms | 0 - 21 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | QCS9075 (Proxy) | QCS9075 Proxy | QNN | 3.898 ms | 1 - 11 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 3.134 ms | 0 - 40 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 3.869 ms | 1 - 38 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 3.122 ms | 1 - 1 MB | INT8 | NPU | -- | | Yolo-v7-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 5.847 ms | 6 - 6 MB | INT8 | NPU | -- | ## License * The license for the original implementation of Yolo-v7-Quantized can be found [here](https://github.com/WongKinYiu/yolov7/blob/main/LICENSE.md). * The license for the compiled assets for on-device deployment can be found [here](https://github.com/WongKinYiu/yolov7/blob/main/LICENSE.md) ## References * [YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors](https://arxiv.org/abs/2207.02696) * [Source Model Implementation](https://github.com/WongKinYiu/yolov7/) ## Community * Join [our AI Hub Slack community](https://qualcomm-ai-hub.slack.com/join/shared_invite/zt-2d5zsmas3-Sj0Q9TzslueCjS31eXG2UA#/shared-invite/email) to collaborate, post questions and learn more about on-device AI. * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com). ## Usage and Limitations Model may not be used for or in connection with any of the following applications: - Accessing essential private and public services and benefits; - Administration of justice and democratic processes; - Assessing or recognizing the emotional state of a person; - Biometric and biometrics-based systems, including categorization of persons based on sensitive characteristics; - Education and vocational training; - Employment and workers management; - Exploitation of the vulnerabilities of persons resulting in harmful behavior; - General purpose social scoring; - Law enforcement; - Management and operation of critical infrastructure; - Migration, asylum and border control management; - Predictive policing; - Real-time remote biometric identification in public spaces; - Recommender systems of social media platforms; - Scraping of facial images (from the internet or otherwise); and/or - Subliminal manipulation