YAML Metadata Warning:The pipeline tag "video-object-tracking" is not in the official list: text-classification, token-classification, table-question-answering, question-answering, zero-shot-classification, translation, summarization, feature-extraction, text-generation, fill-mask, sentence-similarity, text-to-speech, text-to-audio, automatic-speech-recognition, audio-to-audio, audio-classification, audio-text-to-text, voice-activity-detection, depth-estimation, image-classification, object-detection, image-segmentation, text-to-image, image-to-text, image-to-image, image-to-video, unconditional-image-generation, video-classification, reinforcement-learning, robotics, tabular-classification, tabular-regression, tabular-to-text, table-to-text, multiple-choice, text-ranking, text-retrieval, time-series-forecasting, text-to-video, image-text-to-text, image-text-to-image, image-text-to-video, visual-question-answering, document-question-answering, zero-shot-image-classification, graph-ml, mask-generation, zero-shot-object-detection, text-to-3d, image-to-3d, image-feature-extraction, video-text-to-text, keypoint-detection, visual-document-retrieval, any-to-any, video-to-video, other
Track-Anything: Optimized for Qualcomm Devices
Track-Anything is a video based machine learning model to track an object in a video.
This is based on the implementation of Track-Anything 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 2.50, ONNX Runtime 1.27.1 | Download |
| TFLITE | float | Universal | QAIRT 2.50 | Download |
For more device-specific assets and performance metrics, visit Track-Anything 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 Track-Anything on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.video_object_tracking
Model Stats:
- Input resolution: 320x568
- Model checkpoint: xmem
- Model size (encode_key_with_shrinkage) (float): 37.1 MB
- Model size (encode_key_without_shrinkage) (float): 37.1 MB
- Model size (encode_value) (float): 88.8 MB
- Model size (segment) (float): 115 MB
- Number of parameters (encode_key_with_shrinkage): 9.72M
- Number of parameters (encode_key_without_shrinkage): 9.71M
- Number of parameters (encode_value): 23.3M
- Number of parameters (segment): 30.1M
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| encode_key_with_shrinkage | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.385 ms | 1 - 167 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.881 ms | 0 - 163 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Snapdragon® X2 Elite | 2.756 ms | 10 - 10 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Snapdragon® X Elite | 5.196 ms | 18 - 18 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 3.589 ms | 0 - 193 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 8.743 ms | 9 - 186 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 9.419 ms | 8 - 12 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.108 ms | 0 - 20 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Qualcomm® QCS8450 | 8.743 ms | 9 - 186 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 8.628 ms | 9 - 12 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 5.196 ms | 18 - 18 MB | NPU |
| encode_key_with_shrinkage | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 2.881 ms | 0 - 163 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.539 ms | 0 - 179 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.282 ms | 0 - 177 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 4.142 ms | 1 - 208 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 9.501 ms | 3 - 194 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 9.798 ms | 3 - 31 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 6.101 ms | 3 - 5 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Qualcomm® SA8650P | 9.452 ms | 3 - 180 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Qualcomm® SA8255P | 9.452 ms | 3 - 180 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Qualcomm® QCS8450 | 9.501 ms | 3 - 194 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 9.536 ms | 2 - 29 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 3.282 ms | 0 - 177 MB | NPU |
| encode_key_with_shrinkage | TFLITE | float | Qualcomm® SA8295P | 9.96 ms | 3 - 158 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 3.451 ms | 0 - 169 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.407 ms | 51 - 219 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Snapdragon® X2 Elite | 3.874 ms | 80 - 80 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Snapdragon® X Elite | 6.5 ms | 80 - 80 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 4.816 ms | 54 - 249 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 11.088 ms | 50 - 231 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 11.305 ms | 49 - 53 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 6.636 ms | 35 - 54 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Qualcomm® QCS8450 | 11.088 ms | 50 - 231 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 10.442 ms | 51 - 54 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 6.5 ms | 80 - 80 MB | NPU |
| encode_key_without_shrinkage | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 4.407 ms | 51 - 219 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 3.601 ms | 0 - 186 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.364 ms | 20 - 200 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 5.389 ms | 0 - 210 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 12.164 ms | 20 - 216 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 11.785 ms | 20 - 65 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 7.458 ms | 20 - 22 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Qualcomm® SA8650P | 11.663 ms | 20 - 207 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Qualcomm® SA8255P | 11.663 ms | 20 - 207 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Qualcomm® QCS8450 | 12.164 ms | 20 - 216 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 11.257 ms | 20 - 64 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 4.364 ms | 20 - 200 MB | NPU |
| encode_key_without_shrinkage | TFLITE | float | Qualcomm® SA8295P | 12.512 ms | 20 - 177 MB | NPU |
| encode_value | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 5.774 ms | 7 - 174 MB | NPU |
| encode_value | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 7.374 ms | 7 - 173 MB | NPU |
| encode_value | ONNX | float | Snapdragon® X2 Elite | 5.404 ms | 20 - 20 MB | NPU |
| encode_value | ONNX | float | Snapdragon® X Elite | 12.329 ms | 39 - 39 MB | NPU |
| encode_value | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 9.497 ms | 0 - 188 MB | NPU |
| encode_value | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 21.75 ms | 1 - 184 MB | NPU |
| encode_value | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 17.115 ms | 6 - 16 MB | NPU |
| encode_value | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 15.386 ms | 6 - 14 MB | NPU |
| encode_value | ONNX | float | Qualcomm® QCS8450 | 21.75 ms | 1 - 184 MB | NPU |
| encode_value | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 16.973 ms | 6 - 15 MB | NPU |
| encode_value | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 12.329 ms | 39 - 39 MB | NPU |
| encode_value | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 7.374 ms | 7 - 173 MB | NPU |
| encode_value | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 5.088 ms | 1 - 176 MB | NPU |
| encode_value | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.391 ms | 3 - 178 MB | NPU |
| encode_value | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 8.167 ms | 3 - 246 MB | NPU |
| encode_value | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 19.545 ms | 3 - 242 MB | NPU |
| encode_value | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 17.619 ms | 0 - 56 MB | NPU |
| encode_value | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 11.686 ms | 3 - 5 MB | NPU |
| encode_value | TFLITE | float | Qualcomm® SA8650P | 16.996 ms | 3 - 215 MB | NPU |
| encode_value | TFLITE | float | Qualcomm® SA8255P | 16.996 ms | 3 - 215 MB | NPU |
| encode_value | TFLITE | float | Qualcomm® QCS8450 | 19.545 ms | 3 - 242 MB | NPU |
| encode_value | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 16.902 ms | 0 - 55 MB | NPU |
| encode_value | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 6.391 ms | 3 - 178 MB | NPU |
| encode_value | TFLITE | float | Qualcomm® SA8295P | 18.213 ms | 3 - 177 MB | NPU |
| segment | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 9.786 ms | 3 - 204 MB | NPU |
| segment | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 12.775 ms | 3 - 200 MB | NPU |
| segment | ONNX | float | Snapdragon® X2 Elite | 10.382 ms | 21 - 21 MB | NPU |
| segment | ONNX | float | Snapdragon® X Elite | 23.183 ms | 38 - 38 MB | NPU |
| segment | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 17.031 ms | 0 - 232 MB | NPU |
| segment | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 38.068 ms | 16 - 231 MB | NPU |
| segment | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 36.117 ms | 21 - 46 MB | NPU |
| segment | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 23.292 ms | 0 - 44 MB | NPU |
| segment | ONNX | float | Qualcomm® QCS8450 | 38.068 ms | 16 - 231 MB | NPU |
| segment | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 33.774 ms | 21 - 46 MB | NPU |
| segment | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 23.183 ms | 38 - 38 MB | NPU |
| segment | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 12.775 ms | 3 - 200 MB | NPU |
| segment | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 9.909 ms | 0 - 185 MB | NPU |
| segment | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 12.792 ms | 0 - 182 MB | NPU |
| segment | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 17.174 ms | 2 - 273 MB | NPU |
| segment | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 39.199 ms | 2 - 279 MB | NPU |
| segment | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 35.767 ms | 0 - 84 MB | NPU |
| segment | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 22.333 ms | 2 - 4 MB | NPU |
| segment | TFLITE | float | Qualcomm® SA8650P | 33.305 ms | 2 - 194 MB | NPU |
| segment | TFLITE | float | Qualcomm® SA8255P | 33.305 ms | 2 - 194 MB | NPU |
| segment | TFLITE | float | Qualcomm® QCS8450 | 39.199 ms | 2 - 279 MB | NPU |
| segment | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 33.562 ms | 0 - 83 MB | NPU |
| segment | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 12.792 ms | 0 - 182 MB | NPU |
| segment | TFLITE | float | Qualcomm® SA8295P | 36.526 ms | 2 - 187 MB | NPU |
License
- The license for the original implementation of Track-Anything 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.
