--- library_name: pytorch license: other tags: - android pipeline_tag: image-to-image --- ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/ddcolor/web-assets/model_demo.png) # DDColor: Optimized for Qualcomm Devices DDColor is a coloring algorithm that produces natural, vivid color results from incoming black and white images. This is based on the implementation of DDColor found [here](https://github.com/piddnad/DDColor/). 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/v0.59.0/src/qai_hub_models/models/ddcolor) 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 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/ddcolor/releases/v0.59.0/ddcolor-onnx-float.zip) | QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/ddcolor/releases/v0.59.0/ddcolor-qnn_dlc-float.zip) | TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/ddcolor/releases/v0.59.0/ddcolor-tflite-float.zip) For more device-specific assets and performance metrics, visit **[DDColor on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/ddcolor)**. ### Option 2: Export with Custom Configurations Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.59.0/src/qai_hub_models/models/ddcolor) 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 [DDColor on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.59.0/src/qai_hub_models/models/ddcolor) for usage instructions. ## Model Details **Model Type:** Model_use_case.image_editing **Model Stats:** - Model checkpoint: ddcolor_paper_tiny.pth - Input resolution: 224x224 - Number of parameters: 56.3M - Model size (float): 215 MB - Model size (w8a8): 54.8 MB ## Performance Summary | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |---|---|---|---|---|---|--- | DDColor | ONNX | float | Snapdragon® X2 Elite | 28.048 ms | 2 - 2 MB | NPU | DDColor | ONNX | float | Snapdragon® X Elite | 72.597 ms | 112 - 112 MB | NPU | DDColor | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 44.686 ms | 2 - 1476 MB | NPU | DDColor | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 104.091 ms | 3 - 538 MB | NPU | DDColor | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 74.362 ms | 0 - 540 MB | NPU | DDColor | ONNX | float | Qualcomm® QCS8450 | 104.091 ms | 3 - 538 MB | NPU | DDColor | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 80.439 ms | 1 - 4 MB | NPU | DDColor | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 72.597 ms | 112 - 112 MB | NPU | DDColor | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 35.203 ms | 2 - 801 MB | NPU | DDColor | ONNX | float | Snapdragon® 8 Elite Mobile | 35.203 ms | 2 - 801 MB | NPU | DDColor | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 27.458 ms | 2 - 740 MB | NPU | DDColor | QNN_DLC | float | Snapdragon® X2 Elite | 27.744 ms | 1 - 1 MB | NPU | DDColor | QNN_DLC | float | Snapdragon® X Elite | 62.37 ms | 1 - 1 MB | NPU | DDColor | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 41.678 ms | 0 - 559 MB | NPU | DDColor | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 73.513 ms | 0 - 462 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8275 | 129.721 ms | 1 - 453 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 61.812 ms | 1 - 456 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® SA8775P | 65.654 ms | 1 - 453 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® SA8650P | 65.654 ms | 1 - 453 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® SA8255P | 65.654 ms | 1 - 453 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® QCS8450 | 73.513 ms | 0 - 462 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 66.151 ms | 1 - 4 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 62.37 ms | 1 - 1 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 31.297 ms | 1 - 620 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® SA7255P | 129.721 ms | 1 - 453 MB | NPU | DDColor | QNN_DLC | float | Qualcomm® SA8295P | 69.55 ms | 0 - 330 MB | NPU | DDColor | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 31.297 ms | 1 - 620 MB | NPU | DDColor | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 27.368 ms | 0 - 642 MB | NPU | DDColor | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 42.944 ms | 1 - 1426 MB | NPU | DDColor | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 81.11 ms | 1 - 554 MB | NPU | DDColor | TFLITE | float | Qualcomm® Dragonwing™ QCS8275 | 124.433 ms | 1 - 802 MB | NPU | DDColor | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 67.966 ms | 1 - 4 MB | NPU | DDColor | TFLITE | float | Qualcomm® SA8775P | 72.388 ms | 1 - 851 MB | NPU | DDColor | TFLITE | float | Qualcomm® SA8650P | 72.388 ms | 1 - 851 MB | NPU | DDColor | TFLITE | float | Qualcomm® SA8255P | 72.388 ms | 1 - 851 MB | NPU | DDColor | TFLITE | float | Qualcomm® QCS8450 | 81.11 ms | 1 - 554 MB | NPU | DDColor | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 72.594 ms | 1 - 116 MB | NPU | DDColor | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 33.701 ms | 1 - 978 MB | NPU | DDColor | TFLITE | float | Qualcomm® SA7255P | 124.433 ms | 1 - 802 MB | NPU | DDColor | TFLITE | float | Qualcomm® SA8295P | 73.625 ms | 1 - 378 MB | NPU | DDColor | TFLITE | float | Snapdragon® 8 Elite Mobile | 33.701 ms | 1 - 978 MB | NPU | DDColor | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 27.498 ms | 1 - 1042 MB | NPU ## License * The license for the original implementation of DDColor can be found [here](https://github.com/piddnad/DDColor/blob/master/LICENSE). ## References * [DDColor: Towards Photo-Realistic Image Colorization via Dual Decoders](https://arxiv.org/abs/2201.03545) * [Source Model Implementation](https://github.com/piddnad/DDColor/) ## 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).