library_name: pytorch
license: other
tags:
- generative_ai
- android
pipeline_tag: unconditional-image-generation
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 |
|---|---|---|---|---|
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X2 Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | 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® X2 Elite | 3.724 ms | 1 - 1 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | 7.589 ms | 418 - 418 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 5.819 ms | 0 - 13 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 7.844 ms | 0 - 4 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 8.104 ms | 0 - 450 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 8.066 ms | 0 - 3 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 7.589 ms | 418 - 418 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 68.805 ms | 0 - 11 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 12.181 ms | 0 - 9 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 4.944 ms | 0 - 12 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 4.944 ms | 0 - 12 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 4.852 ms | 0 - 12 MB | NPU |
| text_encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 12.181 ms | 0 - 9 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | 4.228 ms | 0 - 0 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | 7.853 ms | 0 - 0 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | 5.765 ms | 0 - 8 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 17.398 ms | 0 - 9 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 8.108 ms | 0 - 1 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | 8.327 ms | 0 - 9 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8650P | 8.327 ms | 0 - 9 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8255P | 8.327 ms | 0 - 9 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 9.132 ms | 0 - 2 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 7.853 ms | 0 - 0 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 68.542 ms | 0 - 7 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 12.187 ms | 0 - 6 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 4.952 ms | 0 - 13 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | 17.398 ms | 0 - 9 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | 4.952 ms | 0 - 13 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 4.862 ms | 0 - 9 MB | NPU |
| text_encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | 12.187 ms | 0 - 6 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X2 Elite | 67.102 ms | 1 - 1 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | 157.436 ms | 842 - 842 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 109.461 ms | 0 - 10 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 145.103 ms | 0 - 4 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 152.954 ms | 0 - 898 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 147.514 ms | 0 - 3 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 157.436 ms | 842 - 842 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 2482.97 ms | 1 - 10 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 210.727 ms | 1 - 11 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 102.152 ms | 0 - 12 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 102.152 ms | 0 - 12 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 67.218 ms | 0 - 12 MB | NPU |
| unet | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 210.727 ms | 1 - 11 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | 67.816 ms | 0 - 0 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | 157.776 ms | 1 - 1 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | 109.099 ms | 3 - 10 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 346.234 ms | 0 - 10 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 152.677 ms | 0 - 2 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | 147.817 ms | 0 - 9 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8650P | 147.817 ms | 0 - 9 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8255P | 147.817 ms | 0 - 9 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 146.509 ms | 0 - 2 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 157.776 ms | 1 - 1 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 2374.723 ms | 0 - 6 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 211.818 ms | 0 - 7 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 98.588 ms | 1 - 13 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | 346.234 ms | 0 - 10 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | 98.588 ms | 1 - 13 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 65.846 ms | 1 - 9 MB | NPU |
| unet | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | 211.818 ms | 0 - 7 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X2 Elite | 128.514 ms | 4 - 4 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | 257.883 ms | 65 - 65 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 197.505 ms | 5 - 17 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 261.974 ms | 3 - 7 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 263.82 ms | 3 - 5 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 241.322 ms | 3 - 6 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 257.883 ms | 65 - 65 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 6291.567 ms | 3 - 9 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 503.029 ms | 2 - 8 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 160.713 ms | 4 - 15 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 160.713 ms | 4 - 15 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 120.718 ms | 3 - 10 MB | NPU |
| vae | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 503.029 ms | 2 - 8 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | 128.832 ms | 0 - 0 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | 258.113 ms | 0 - 0 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | 196.962 ms | 1 - 8 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 713.323 ms | 0 - 9 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 263.877 ms | 1 - 3 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | 244.569 ms | 0 - 10 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8650P | 244.569 ms | 0 - 10 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8255P | 244.569 ms | 0 - 10 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 241.406 ms | 0 - 3 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 258.113 ms | 0 - 0 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 6235.794 ms | 3 - 10 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 503.975 ms | 0 - 6 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 160.055 ms | 1 - 13 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | 713.323 ms | 0 - 9 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | 160.055 ms | 1 - 13 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 121.283 ms | 0 - 9 MB | NPU |
| vae | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | 503.975 ms | 0 - 6 MB | NPU |
License
- The license for the original implementation of Stable-Diffusion-v2.1 can be found here.
References
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
