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. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.

Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up 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
QNN_DLC	float	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit DDColor on Qualcomm® AI Hub.

Option 2: Export with Custom Configurations

Use the Qualcomm® AI Hub Models 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 for usage instructions.

Model Details

Model Type: Model_use_case.image_editing

Model Stats:

• Input resolution: 224x224
• Model checkpoint: ddcolor_paper_tiny.pth
• Model size (float): 215 MB
• Model size (w8a8): 54.8 MB
• Number of parameters: 56.3M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
DDColor	ONNX	float	Snapdragon® X2 Elite	28.205 ms	2 - 2 MB	NPU
DDColor	ONNX	float	Snapdragon® X Elite	72.412 ms	112 - 112 MB	NPU
DDColor	ONNX	float	Snapdragon® 8 Gen 3 Mobile	46.458 ms	2 - 1484 MB	NPU
DDColor	ONNX	float	Snapdragon® 8 Gen 1 Mobile	85.859 ms	2 - 637 MB	NPU
DDColor	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	63.169 ms	1 - 5 MB	NPU
DDColor	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	73.064 ms	0 - 130 MB	NPU
DDColor	ONNX	float	Qualcomm® QCS8450	85.859 ms	2 - 637 MB	NPU
DDColor	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	81.613 ms	1 - 4 MB	NPU
DDColor	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	72.412 ms	112 - 112 MB	NPU
DDColor	ONNX	float	Qualcomm® Dragonwing™ Q-8750	35.256 ms	1 - 798 MB	NPU
DDColor	ONNX	float	Snapdragon® 8 Elite Mobile	35.256 ms	1 - 798 MB	NPU
DDColor	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	27.415 ms	2 - 739 MB	NPU
DDColor	QNN_DLC	float	Snapdragon® X2 Elite	27.758 ms	1 - 1 MB	NPU
DDColor	QNN_DLC	float	Snapdragon® X Elite	62.449 ms	1 - 1 MB	NPU
DDColor	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	41.563 ms	1 - 560 MB	NPU
DDColor	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	73.31 ms	0 - 462 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	59.598 ms	1 - 4 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	62.498 ms	1 - 11 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® SA8775P	65.734 ms	1 - 452 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® SA8650P	65.734 ms	1 - 452 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® SA8255P	65.734 ms	1 - 452 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® QCS8450	73.31 ms	0 - 462 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	66.518 ms	1 - 4 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	62.449 ms	1 - 1 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	31.624 ms	0 - 615 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® SA7255P	129.606 ms	1 - 453 MB	NPU
DDColor	QNN_DLC	float	Qualcomm® SA8295P	69.507 ms	0 - 330 MB	NPU
DDColor	QNN_DLC	float	Snapdragon® 8 Elite Mobile	31.624 ms	0 - 615 MB	NPU
DDColor	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	27.375 ms	0 - 634 MB	NPU
DDColor	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	43.13 ms	1 - 1422 MB	NPU
DDColor	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	80.963 ms	1 - 550 MB	NPU
DDColor	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	58.147 ms	1 - 117 MB	NPU
DDColor	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	67.849 ms	1 - 4 MB	NPU
DDColor	TFLITE	float	Qualcomm® SA8775P	72.46 ms	1 - 852 MB	NPU
DDColor	TFLITE	float	Qualcomm® SA8650P	72.46 ms	1 - 852 MB	NPU
DDColor	TFLITE	float	Qualcomm® SA8255P	72.46 ms	1 - 852 MB	NPU
DDColor	TFLITE	float	Qualcomm® QCS8450	80.963 ms	1 - 550 MB	NPU
DDColor	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	71.896 ms	1 - 116 MB	NPU
DDColor	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	33.823 ms	1 - 979 MB	NPU
DDColor	TFLITE	float	Qualcomm® SA7255P	124.44 ms	1 - 801 MB	NPU
DDColor	TFLITE	float	Qualcomm® SA8295P	73.57 ms	1 - 378 MB	NPU
DDColor	TFLITE	float	Snapdragon® 8 Elite Mobile	33.823 ms	1 - 979 MB	NPU
DDColor	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	27.512 ms	1 - 1037 MB	NPU

License

• The license for the original implementation of DDColor can be found here.

References

• DDColor: Towards Photo-Realistic Image Colorization via Dual Decoders
• Source Model Implementation

Community

• Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
• For questions or feedback please reach out to us.