v0.49.1
Browse filesSee https://github.com/qualcomm/ai-hub-models/releases/v0.49.1 for changelog.
README.md
CHANGED
|
@@ -14,7 +14,7 @@ pipeline_tag: object-detection
|
|
| 14 |
CenterNet-2D is machine learning model that detects objects by finding their center points.
|
| 15 |
|
| 16 |
This is based on the implementation of CenterNet-2D found [here](https://github.com/xingyizhou/CenterNet).
|
| 17 |
-
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/
|
| 18 |
|
| 19 |
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.
|
| 20 |
|
|
@@ -27,38 +27,38 @@ Below are pre-exported model assets ready for deployment.
|
|
| 27 |
|
| 28 |
| Runtime | Precision | Chipset | SDK Versions | Download |
|
| 29 |
|---|---|---|---|---|
|
| 30 |
-
| PRECOMPILED_QNN_ONNX | float | Snapdragon®
|
| 31 |
-
| PRECOMPILED_QNN_ONNX | float | Snapdragon®
|
| 32 |
-
| PRECOMPILED_QNN_ONNX | float | Snapdragon®
|
| 33 |
-
| PRECOMPILED_QNN_ONNX | float |
|
| 34 |
-
| PRECOMPILED_QNN_ONNX | float |
|
| 35 |
-
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite
|
| 36 |
-
| PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS9075 | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.
|
| 37 |
-
| QNN_CONTEXT_BINARY | float | Snapdragon®
|
| 38 |
-
| QNN_CONTEXT_BINARY | float | Snapdragon®
|
| 39 |
-
| QNN_CONTEXT_BINARY | float | Snapdragon®
|
| 40 |
-
| QNN_CONTEXT_BINARY | float |
|
| 41 |
-
| QNN_CONTEXT_BINARY | float | Qualcomm®
|
| 42 |
-
| QNN_CONTEXT_BINARY | float |
|
| 43 |
-
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite
|
| 44 |
-
| QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.
|
| 45 |
-
| QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.
|
| 46 |
-
| QNN_CONTEXT_BINARY | float | Qualcomm® QCS9075 | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.
|
| 47 |
-
| QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 (Proxy) | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.
|
| 48 |
|
| 49 |
For more device-specific assets and performance metrics, visit **[CenterNet-2D on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/centernet_2d)**.
|
| 50 |
|
| 51 |
|
| 52 |
### Option 2: Export with Custom Configurations
|
| 53 |
|
| 54 |
-
Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/
|
| 55 |
- Custom weights (e.g., fine-tuned checkpoints)
|
| 56 |
- Custom input shapes
|
| 57 |
- Target device and runtime configurations
|
| 58 |
|
| 59 |
This option is ideal if you need to customize the model beyond the default configuration provided here.
|
| 60 |
|
| 61 |
-
See our repository for [CenterNet-2D on GitHub](https://github.com/qualcomm/ai-hub-models/
|
| 62 |
|
| 63 |
## Model Details
|
| 64 |
|
|
@@ -73,25 +73,25 @@ See our repository for [CenterNet-2D on GitHub](https://github.com/qualcomm/ai-h
|
|
| 73 |
## Performance Summary
|
| 74 |
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|
| 75 |
|---|---|---|---|---|---|---
|
| 76 |
-
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon®
|
| 77 |
-
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon®
|
| 78 |
-
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon®
|
| 79 |
-
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float |
|
| 80 |
-
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Qualcomm®
|
| 81 |
-
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float |
|
| 82 |
-
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite
|
| 83 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon®
|
| 84 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon®
|
| 85 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon®
|
| 86 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float |
|
| 87 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm®
|
| 88 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm®
|
| 89 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm®
|
| 90 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm®
|
| 91 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm®
|
| 92 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm®
|
| 93 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float |
|
| 94 |
-
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite
|
| 95 |
|
| 96 |
## License
|
| 97 |
* The license for the original implementation of CenterNet-2D can be found
|
|
|
|
| 14 |
CenterNet-2D is machine learning model that detects objects by finding their center points.
|
| 15 |
|
| 16 |
This is based on the implementation of CenterNet-2D found [here](https://github.com/xingyizhou/CenterNet).
|
| 17 |
+
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/tree/v0.49.1/qai_hub_models/models/centernet_2d) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
|
| 18 |
|
| 19 |
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.
|
| 20 |
|
|
|
|
| 27 |
|
| 28 |
| Runtime | Precision | Chipset | SDK Versions | Download |
|
| 29 |
|---|---|---|---|---|
|
| 30 |
+
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5.zip)
|
| 31 |
+
| PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip)
|
| 32 |
+
| PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip)
|
| 33 |
+
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip)
|
| 34 |
+
| PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8550 (Proxy) | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip)
|
| 35 |
+
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
|
| 36 |
+
| PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS9075 | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip)
|
| 37 |
+
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5.zip)
|
| 38 |
+
| QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip)
|
| 39 |
+
| QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip)
|
| 40 |
+
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip)
|
| 41 |
+
| QNN_CONTEXT_BINARY | float | Qualcomm® QCS8550 (Proxy) | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip)
|
| 42 |
+
| QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_sa8775p.zip)
|
| 43 |
+
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
|
| 44 |
+
| QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_sa7255p.zip)
|
| 45 |
+
| QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_sa8295p.zip)
|
| 46 |
+
| QNN_CONTEXT_BINARY | float | Qualcomm® QCS9075 | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_qcs9075.zip)
|
| 47 |
+
| QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 (Proxy) | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/centernet_2d/releases/v0.49.1/centernet_2d-qnn_context_binary-float-qualcomm_qcs8450_proxy.zip)
|
| 48 |
|
| 49 |
For more device-specific assets and performance metrics, visit **[CenterNet-2D on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/centernet_2d)**.
|
| 50 |
|
| 51 |
|
| 52 |
### Option 2: Export with Custom Configurations
|
| 53 |
|
| 54 |
+
Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/tree/v0.49.1/qai_hub_models/models/centernet_2d) Python library to compile and export the model with your own:
|
| 55 |
- Custom weights (e.g., fine-tuned checkpoints)
|
| 56 |
- Custom input shapes
|
| 57 |
- Target device and runtime configurations
|
| 58 |
|
| 59 |
This option is ideal if you need to customize the model beyond the default configuration provided here.
|
| 60 |
|
| 61 |
+
See our repository for [CenterNet-2D on GitHub](https://github.com/qualcomm/ai-hub-models/tree/v0.49.1/qai_hub_models/models/centernet_2d) for usage instructions.
|
| 62 |
|
| 63 |
## Model Details
|
| 64 |
|
|
|
|
| 73 |
## Performance Summary
|
| 74 |
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|
| 75 |
|---|---|---|---|---|---|---
|
| 76 |
+
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 241.51 ms | 17 - 26 MB | NPU
|
| 77 |
+
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 267.704 ms | 52 - 52 MB | NPU
|
| 78 |
+
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 460.776 ms | 54 - 54 MB | NPU
|
| 79 |
+
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 307.314 ms | 16 - 22 MB | NPU
|
| 80 |
+
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8550 (Proxy) | 449.89 ms | 1 - 62 MB | NPU
|
| 81 |
+
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS9075 | 462.495 ms | 9 - 15 MB | NPU
|
| 82 |
+
| CenterNet-2D | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 310.884 ms | 15 - 23 MB | NPU
|
| 83 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | 242.21 ms | 3 - 12 MB | NPU
|
| 84 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 253.484 ms | 3 - 3 MB | NPU
|
| 85 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 442.752 ms | 3 - 3 MB | NPU
|
| 86 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 320.865 ms | 4 - 11 MB | NPU
|
| 87 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8275 (Proxy) | 583.12 ms | 0 - 9 MB | NPU
|
| 88 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8550 (Proxy) | 439.373 ms | 3 - 4 MB | NPU
|
| 89 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | 477.609 ms | 2 - 10 MB | NPU
|
| 90 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm® QCS9075 | 468.103 ms | 3 - 13 MB | NPU
|
| 91 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 (Proxy) | 597.271 ms | 3 - 12 MB | NPU
|
| 92 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | 583.12 ms | 0 - 9 MB | NPU
|
| 93 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 497.803 ms | 0 - 5 MB | NPU
|
| 94 |
+
| CenterNet-2D | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite For Galaxy Mobile | 306.697 ms | 0 - 9 MB | NPU
|
| 95 |
|
| 96 |
## License
|
| 97 |
* The license for the original implementation of CenterNet-2D can be found
|