InternImage: Optimized for Qualcomm Devices
InternImage employs DCNv3 as its core operator to equips the model with dynamic and effective receptive fields required for downstream tasks like object detection and segmentation, while enabling adaptive spatial aggregation.
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 |
|---|---|---|---|---|
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8550 (Proxy) | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS9075 | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Snapdragon® X2 Elite | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Snapdragon® X Elite | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Qualcomm® QCS6490 | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Qualcomm® QCM6690 | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a8 | Qualcomm® QCS9075 | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® QCS8550 (Proxy) | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® QCS9075 | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 (Proxy) | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Snapdragon® X2 Elite | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Snapdragon® X Elite | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Qualcomm® QCS6490 | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Qualcomm® QCS8550 (Proxy) | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Qualcomm® SA8775P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Qualcomm® SA7255P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Qualcomm® SA8295P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Qualcomm® QCM6690 | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Qualcomm® QCS9075 | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | w8a8 | Qualcomm® QCS8450 (Proxy) | QAIRT 2.43 | Download |
For more device-specific assets and performance metrics, visit InternImage 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 InternImage on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.image_classification
Model Stats:
- Model checkpoint: internimage_t_1k_224
- Input resolution: 1x3x224x224
- Number of parameters: 30.6M
- Model size (float): 117 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| InternImage | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 19.914 ms | 1 - 11 MB | NPU |
| InternImage | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 20.716 ms | 66 - 66 MB | NPU |
| InternImage | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 52.11 ms | 66 - 66 MB | NPU |
| InternImage | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 35.042 ms | 1 - 8 MB | NPU |
| InternImage | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8550 (Proxy) | 49.971 ms | 0 - 81 MB | NPU |
| InternImage | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS9075 | 51.979 ms | 0 - 4 MB | NPU |
| InternImage | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 25.49 ms | 1 - 9 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | 20.36 ms | 1 - 10 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 21.446 ms | 1 - 1 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 53.328 ms | 1 - 1 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 36.225 ms | 1 - 8 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8275 (Proxy) | 98.738 ms | 1 - 8 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8550 (Proxy) | 51.202 ms | 1 - 3 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Qualcomm® QCS9075 | 52.724 ms | 1 - 3 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 (Proxy) | 59.966 ms | 1 - 9 MB | NPU |
| InternImage | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite For Galaxy Mobile | 25.782 ms | 1 - 9 MB | NPU |
License
- The license for the original implementation of InternImage can be found here.
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.