--- library_name: pytorch license: other tags: - backbone - android pipeline_tag: video-classification --- ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/resnet_mixed/web-assets/model_demo.png) # ResNet-Mixed-Convolution: Optimized for Qualcomm Devices ResNet Mixed Convolutions is a network with a mixture of 2D and 3D convolutions used for video understanding. This is based on the implementation of ResNet-Mixed-Convolution found [here](https://github.com/pytorch/vision/blob/main/torchvision/models/video/resnet.py). This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/resnet_mixed) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary). Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) 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](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/resnet_mixed/releases/v0.52.0/resnet_mixed-onnx-float.zip) | ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.3 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/resnet_mixed/releases/v0.52.0/resnet_mixed-onnx-w8a16.zip) | QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/resnet_mixed/releases/v0.52.0/resnet_mixed-qnn_dlc-float.zip) | QNN_DLC | w8a16 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/resnet_mixed/releases/v0.52.0/resnet_mixed-qnn_dlc-w8a16.zip) | TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/resnet_mixed/releases/v0.52.0/resnet_mixed-tflite-float.zip) For more device-specific assets and performance metrics, visit **[ResNet-Mixed-Convolution on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/resnet_mixed)**. ### Option 2: Export with Custom Configurations Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/resnet_mixed) 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 [ResNet-Mixed-Convolution on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/resnet_mixed) for usage instructions. ## Model Details **Model Type:** Model_use_case.video_classification **Model Stats:** - Model checkpoint: Kinetics-400 - Input resolution: 112x112 - Number of parameters: 11.7M - Model size (float): 44.6 MB - Model size (w8a16): 11.5 MB ## Performance Summary | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |---|---|---|---|---|---|--- | ResNet-Mixed-Convolution | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 5.555 ms | 2 - 221 MB | NPU | ResNet-Mixed-Convolution | ONNX | float | Snapdragon® X2 Elite | 7.095 ms | 22 - 22 MB | NPU | ResNet-Mixed-Convolution | ONNX | float | Snapdragon® X Elite | 13.831 ms | 22 - 22 MB | NPU | ResNet-Mixed-Convolution | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 9.603 ms | 2 - 271 MB | NPU | ResNet-Mixed-Convolution | ONNX | float | Qualcomm® QCS8550 (Proxy) | 13.201 ms | 0 - 28 MB | NPU | ResNet-Mixed-Convolution | ONNX | float | Qualcomm® QCS9075 | 26.756 ms | 2 - 5 MB | NPU | ResNet-Mixed-Convolution | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 7.816 ms | 0 - 215 MB | NPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 3.509 ms | 0 - 191 MB | NPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Snapdragon® X2 Elite | 4.636 ms | 12 - 12 MB | NPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Snapdragon® X Elite | 9.077 ms | 11 - 11 MB | NPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 6.366 ms | 1 - 260 MB | NPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Qualcomm® QCS6490 | 1730.881 ms | 45 - 62 MB | CPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 8.539 ms | 0 - 19 MB | NPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Qualcomm® QCS9075 | 9.046 ms | 1 - 4 MB | NPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Qualcomm® QCM6690 | 896.461 ms | 108 - 115 MB | CPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 5.236 ms | 1 - 193 MB | NPU | ResNet-Mixed-Convolution | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 872.002 ms | 113 - 120 MB | CPU | ResNet-Mixed-Convolution | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 5.767 ms | 2 - 232 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Snapdragon® X2 Elite | 7.737 ms | 2 - 2 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Snapdragon® X Elite | 14.109 ms | 2 - 2 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 9.525 ms | 0 - 278 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 97.175 ms | 1 - 222 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 13.328 ms | 2 - 189 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Qualcomm® SA8775P | 25.217 ms | 0 - 223 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Qualcomm® QCS9075 | 27.938 ms | 2 - 6 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 27.793 ms | 0 - 237 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Qualcomm® SA7255P | 97.175 ms | 1 - 222 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Qualcomm® SA8295P | 26.874 ms | 0 - 195 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 7.64 ms | 0 - 226 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 4.034 ms | 1 - 195 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 5.485 ms | 1 - 1 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Snapdragon® X Elite | 10.082 ms | 1 - 1 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 6.769 ms | 0 - 242 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® QCS6490 | 36.957 ms | 3 - 5 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 30.055 ms | 1 - 190 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 9.427 ms | 1 - 169 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® SA8775P | 9.357 ms | 1 - 191 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 10.38 ms | 0 - 3 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 180.867 ms | 1 - 210 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® QCS8450 (Proxy) | 12.773 ms | 1 - 244 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® SA7255P | 30.055 ms | 1 - 190 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Qualcomm® SA8295P | 16.109 ms | 0 - 193 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 5.366 ms | 1 - 188 MB | NPU | ResNet-Mixed-Convolution | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 15.948 ms | 1 - 199 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 206.533 ms | 0 - 256 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 221.782 ms | 0 - 305 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 593.879 ms | 0 - 253 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 305.52 ms | 0 - 2 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Qualcomm® SA8775P | 312.41 ms | 0 - 251 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Qualcomm® QCS9075 | 313.28 ms | 0 - 28 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 341.797 ms | 0 - 268 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Qualcomm® SA7255P | 593.879 ms | 0 - 253 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Qualcomm® SA8295P | 353.483 ms | 0 - 235 MB | NPU | ResNet-Mixed-Convolution | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 201.549 ms | 0 - 256 MB | NPU ## License * The license for the original implementation of ResNet-Mixed-Convolution can be found [here](https://github.com/pytorch/vision/blob/main/LICENSE). ## References * [A Closer Look at Spatiotemporal Convolutions for Action Recognition](https://arxiv.org/abs/1711.11248) * [Source Model Implementation](https://github.com/pytorch/vision/blob/main/torchvision/models/video/resnet.py) ## Community * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) 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).