Stable-Diffusion-v2.1: Optimized for Qualcomm Devices

Generates high resolution images from text prompts using a latent diffusion model. This model uses CLIP ViT-L/14 as text encoder, U-Net based latent denoising, and VAE based decoder to generate the final image.

This is based on the implementation of Stable-Diffusion-v2.1 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.

Deploy to Snapdragon X Elite NPU

Please follow the Stable Diffusion Windows App tutorial to quantize model with custom weights.

Quantize and Deploy Your Own Fine-Tuned Stable Diffusion

Please follow the Quantize Stable Diffusion tutorial to quantize model with custom weights.

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
QNN_CONTEXT_BINARY	w8a16	qualcomm_qcm6690	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_qcs8550_proxy	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_qcs9075	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_sa7255p	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_sa8775p	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_snapdragon_7gen4	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_snapdragon_8_elite_for_galaxy	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_snapdragon_8_elite_gen5	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_snapdragon_8gen3	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_snapdragon_x2_elite	QAIRT 2.43	Download
QNN_CONTEXT_BINARY	w8a16	qualcomm_snapdragon_x_elite	QAIRT 2.43	Download

For more device-specific assets and performance metrics, visit Stable-Diffusion-v2.1 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 Stable-Diffusion-v2.1 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_generation

Model Stats:

• Input: Text prompt to generate image

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	4.618 ms	0 - 10 MB	NPU
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® X2 Elite	3.227 ms	417 - 417 MB	NPU
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® X Elite	7.744 ms	418 - 418 MB	NPU
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	5.954 ms	0 - 11 MB	NPU
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	8.013 ms	0 - 450 MB	NPU
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCS9075	7.387 ms	0 - 3 MB	NPU
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCM6690	81.587 ms	0 - 10 MB	NPU
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	5.061 ms	0 - 10 MB	NPU
text_encoder	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	12.228 ms	0 - 11 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	4.762 ms	0 - 10 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Snapdragon® X2 Elite	3.973 ms	0 - 0 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Snapdragon® X Elite	7.867 ms	0 - 0 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Gen 3 Mobile	5.975 ms	0 - 7 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS8275 (Proxy)	17.415 ms	0 - 8 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS8550 (Proxy)	8.189 ms	0 - 1 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Qualcomm® SA8775P	8.175 ms	0 - 9 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS9075	7.85 ms	0 - 2 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCM6690	68.878 ms	0 - 6 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Qualcomm® SA7255P	17.415 ms	0 - 8 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	4.97 ms	0 - 13 MB	NPU
text_encoder	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 7 Gen 4 Mobile	12.048 ms	0 - 6 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	69.149 ms	0 - 9 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® X2 Elite	65.93 ms	845 - 845 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® X Elite	157.739 ms	842 - 842 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	110.007 ms	0 - 12 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	153.135 ms	0 - 898 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCS9075	148.971 ms	0 - 3 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCM6690	2542.651 ms	0 - 11 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	101.935 ms	1 - 12 MB	NPU
unet	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	218.303 ms	0 - 11 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	65.561 ms	0 - 9 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Snapdragon® X2 Elite	64.358 ms	1 - 1 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Snapdragon® X Elite	156.663 ms	0 - 0 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Gen 3 Mobile	109.689 ms	0 - 8 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS8275 (Proxy)	343.02 ms	0 - 7 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS8550 (Proxy)	152.196 ms	0 - 2 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Qualcomm® SA8775P	146.722 ms	0 - 7 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS9075	146.499 ms	0 - 2 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCM6690	2801.418 ms	1 - 11 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Qualcomm® SA7255P	343.02 ms	0 - 7 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	96.033 ms	1 - 9 MB	NPU
unet	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 7 Gen 4 Mobile	236.036 ms	0 - 6 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	133.06 ms	3 - 13 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® X2 Elite	143.887 ms	63 - 63 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® X Elite	260.237 ms	64 - 64 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	197.699 ms	3 - 10 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	262.72 ms	0 - 66 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCS9075	246.076 ms	3 - 6 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Qualcomm® QCM6690	6256.454 ms	1 - 7 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	171.626 ms	3 - 11 MB	NPU
vae	PRECOMPILED_QNN_ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	507.866 ms	1 - 8 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	123.858 ms	0 - 10 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Snapdragon® X2 Elite	130.434 ms	0 - 0 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Snapdragon® X Elite	256.6 ms	0 - 0 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Gen 3 Mobile	196.357 ms	0 - 7 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS8275 (Proxy)	710.583 ms	0 - 9 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS8550 (Proxy)	260.722 ms	1 - 3 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Qualcomm® SA8775P	246.202 ms	0 - 9 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCS9075	243.611 ms	0 - 3 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Qualcomm® QCM6690	6253.726 ms	0 - 6 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Qualcomm® SA7255P	710.583 ms	0 - 9 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	160.62 ms	0 - 9 MB	NPU
vae	QNN_CONTEXT_BINARY	w8a16	Snapdragon® 7 Gen 4 Mobile	508.249 ms	0 - 7 MB	NPU

License

• The license for the original implementation of Stable-Diffusion-v2.1 can be found here.

References

• High-Resolution Image Synthesis with Latent Diffusion Models
• 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.

Usage and Limitations

This model may not be used for or in connection with any of the following applications:

• Accessing essential private and public services and benefits;
• Administration of justice and democratic processes;
• Assessing or recognizing the emotional state of a person;
• Biometric and biometrics-based systems, including categorization of persons based on sensitive characteristics;
• Education and vocational training;
• Employment and workers management;
• Exploitation of the vulnerabilities of persons resulting in harmful behavior;
• General purpose social scoring;
• Law enforcement;
• Management and operation of critical infrastructure;
• Migration, asylum and border control management;
• Predictive policing;
• Real-time remote biometric identification in public spaces;
• Recommender systems of social media platforms;
• Scraping of facial images (from the internet or otherwise); and/or
• Subliminal manipulation