--- library_name: pytorch license: mit tags: - android pipeline_tag: image-to-video --- ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/fomm/web-assets/model_demo.png) # First-Order-Motion-Model: Optimized for Qualcomm Devices FOMM is a machine learning model that animates a still image to mirror the movements from a target video. This is based on the implementation of First-Order-Motion-Model found [here](https://github.com/AliaksandrSiarohin/first-order-model/tree/master). 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/v0.63.0/src/qai_hub_models/models/fomm) 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 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/fomm/releases/v0.63.0/fomm-onnx-float.zip) For more device-specific assets and performance metrics, visit **[First-Order-Motion-Model on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/fomm)**. ### Option 2: Export with Custom Configurations Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.63.0/src/qai_hub_models/models/fomm) 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 [First-Order-Motion-Model on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.63.0/src/qai_hub_models/models/fomm) for usage instructions. ## Model Details **Model Type:** Model_use_case.video_generation **Model Stats:** - Input resolution: 256x256 - Model checkpoint: vox-256 - Model size (detector) (float): 54.2 MB - Model size (generator) (float): 174 MB ## Performance Summary | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |---|---|---|---|---|---|--- | detector | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.757 ms | 0 - 26 MB | NPU | detector | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.902 ms | 0 - 24 MB | NPU | detector | ONNX | float | Snapdragon® X2 Elite | 2.664 ms | 2 - 2 MB | NPU | detector | ONNX | float | Snapdragon® X Elite | 4.533 ms | 28 - 28 MB | NPU | detector | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 3.291 ms | 0 - 36 MB | NPU | detector | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 7.375 ms | 1 - 42 MB | NPU | detector | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 6.605 ms | 1 - 5 MB | NPU | detector | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 4.42 ms | 0 - 43 MB | NPU | detector | ONNX | float | Qualcomm® QCS8450 | 7.375 ms | 1 - 42 MB | NPU | detector | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 5.643 ms | 1 - 4 MB | NPU | detector | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 4.533 ms | 28 - 28 MB | NPU | detector | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 2.902 ms | 0 - 24 MB | NPU | detector | TFLITE | float | Qualcomm® SA8775P | 5.807 ms | 0 - 29 MB | NPU | detector | TFLITE | float | Qualcomm® SA8650P | 5.807 ms | 0 - 29 MB | NPU | detector | TFLITE | float | Qualcomm® SA8255P | 5.807 ms | 0 - 29 MB | NPU | detector | TFLITE | float | Qualcomm® SA7255P | 20.3 ms | 0 - 16 MB | GPU | generator | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 10.869 ms | 17 - 184 MB | NPU | generator | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 13.285 ms | 15 - 170 MB | NPU | generator | ONNX | float | Snapdragon® X2 Elite | 12.27 ms | 23 - 23 MB | NPU | generator | ONNX | float | Snapdragon® X Elite | 22.403 ms | 89 - 89 MB | NPU | generator | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 15.94 ms | 0 - 186 MB | NPU | generator | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 35.502 ms | 16 - 199 MB | NPU | generator | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 38.125 ms | 16 - 20 MB | NPU | generator | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 21.94 ms | 18 - 20 MB | NPU | generator | ONNX | float | Qualcomm® QCS8450 | 35.502 ms | 16 - 199 MB | NPU | generator | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 34.003 ms | 17 - 21 MB | NPU | generator | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 22.403 ms | 89 - 89 MB | NPU | generator | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 13.285 ms | 15 - 170 MB | NPU | generator | TFLITE | float | Qualcomm® SA8775P | 545.416 ms | 19 - 37 MB | CPU | generator | TFLITE | float | Qualcomm® SA8650P | 545.416 ms | 19 - 37 MB | CPU | generator | TFLITE | float | Qualcomm® SA8255P | 545.416 ms | 19 - 37 MB | CPU | generator | TFLITE | float | Qualcomm® SA7255P | 2876.088 ms | 21 - 37 MB | CPU ## License * The license for the original implementation of First-Order-Motion-Model can be found [here](https://github.com/AliaksandrSiarohin/first-order-model/blob/master/LICENSE.md). ## References * [First Order Motion Model for Image Animation](https://arxiv.org/abs/2003.00196) * [Source Model Implementation](https://github.com/AliaksandrSiarohin/first-order-model/tree/master) ## 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).