--- library_name: pytorch license: other 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 Mobile Deployment ## Animation of Still Image from Source Video FOMM is a machine learning model that animates a still image to mirror the movements from a target video. This model is an implementation of First-Order-Motion-Model found [here](https://github.com/AliaksandrSiarohin/first-order-model/tree/master). This repository provides scripts to run First-Order-Motion-Model on Qualcomm® devices. More details on model performance across various devices, can be found [here](https://aihub.qualcomm.com/models/fomm). ### Model Details - **Model Type:** Model_use_case.video_generation - **Model Stats:** - Model checkpoint: vox-256 - Input resolution: 256x256 - Model size (FOMMDetector) (float): 54.2 MB - Model size (FOMMGenerator) (float): 174 MB | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model |---|---|---|---|---|---|---|---|---| | FOMMDetector | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 4.798 ms | 0 - 78 MB | NPU | [First-Order-Motion-Model.onnx](https://huggingface.co/qualcomm/First-Order-Motion-Model/blob/main/First-Order-Motion-Model.onnx) | | FOMMDetector | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 3.551 ms | 1 - 21 MB | NPU | [First-Order-Motion-Model.onnx](https://huggingface.co/qualcomm/First-Order-Motion-Model/blob/main/First-Order-Motion-Model.onnx) | | FOMMDetector | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 3.887 ms | 1 - 18 MB | NPU | [First-Order-Motion-Model.onnx](https://huggingface.co/qualcomm/First-Order-Motion-Model/blob/main/First-Order-Motion-Model.onnx) | | FOMMDetector | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 4.987 ms | 29 - 29 MB | NPU | [First-Order-Motion-Model.onnx](https://huggingface.co/qualcomm/First-Order-Motion-Model/blob/main/First-Order-Motion-Model.onnx) | | FOMMGenerator | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 26.686 ms | 0 - 101 MB | NPU | [First-Order-Motion-Model.onnx](https://huggingface.co/qualcomm/First-Order-Motion-Model/blob/main/First-Order-Motion-Model.onnx) | | FOMMGenerator | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 19.034 ms | 23 - 93 MB | NPU | [First-Order-Motion-Model.onnx](https://huggingface.co/qualcomm/First-Order-Motion-Model/blob/main/First-Order-Motion-Model.onnx) | | FOMMGenerator | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 15.988 ms | 20 - 84 MB | NPU | [First-Order-Motion-Model.onnx](https://huggingface.co/qualcomm/First-Order-Motion-Model/blob/main/First-Order-Motion-Model.onnx) | | FOMMGenerator | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 25.059 ms | 88 - 88 MB | NPU | [First-Order-Motion-Model.onnx](https://huggingface.co/qualcomm/First-Order-Motion-Model/blob/main/First-Order-Motion-Model.onnx) | ## Installation Install the package via pip: ```bash pip install "qai-hub-models[fomm]" ``` ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) 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. ```bash qai-hub configure --api_token API_TOKEN ``` Navigate to [docs](https://app.aihub.qualcomm.com/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. ```bash python -m qai_hub_models.models.fomm.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.fomm.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. ```bash python -m qai_hub_models.models.fomm.export ``` ``` Profiling Results ------------------------------------------------------------ FOMMDetector Device : cs_8_gen_2 (ANDROID 13) Runtime : ONNX Estimated inference time (ms) : 4.8 Estimated peak memory usage (MB): [0, 78] Total # Ops : 57 Compute Unit(s) : npu (57 ops) gpu (0 ops) cpu (0 ops) ------------------------------------------------------------ FOMMGenerator Device : cs_8_gen_2 (ANDROID 13) Runtime : ONNX Estimated inference time (ms) : 26.7 Estimated peak memory usage (MB): [0, 101] Total # Ops : 151 Compute Unit(s) : npu (139 ops) gpu (0 ops) cpu (12 ops) ``` ## How does this work? This [export script](https://aihub.qualcomm.com/models/fomm/qai_hub_models/models/First-Order-Motion-Model/export.py) leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) 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. ```python import torch import qai_hub as hub from qai_hub_models.models.fomm import Model # Load the model torch_model = Model.from_pretrained() # Device device = hub.Device("Samsung Galaxy S24") # 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. ```python 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. ```python 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](https://myaccount.qualcomm.com/signup). ## Deploying compiled model to Android The models can be deployed using multiple runtimes: - TensorFlow Lite (`.tflite` export): [This tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a guide to deploy the .tflite model in an Android application. - QNN (`.so` export ): This [sample app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html) provides instructions on how to use the `.so` shared library in an Android application. ## View on Qualcomm® AI Hub Get more details on First-Order-Motion-Model's performance across various devices [here](https://aihub.qualcomm.com/models/fomm). Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/) ## 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). * The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf) ## 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).