Track Anything: Segment Anything Meets Videos
Paper
• 2304.11968 • Published
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 2.42, ONNX Runtime 1.24.1 | Download |
| TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | 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:
- Model checkpoint: xmem
- Input resolution: 320x568
- Number of parameters (TrackAnythingEncodeKeyWithShrinkage): 9.72M
- Model size (TrackAnythingEncodeKeyWithShrinkage) (float): 37.1 MB
- Number of parameters (TrackAnythingEncodeValue): 23.3M
- Model size (TrackAnythingEncodeValue) (float): 88.8 MB
- Number of parameters (TrackAnythingEncodeKeyWithoutShrinkage): 9.71M
- Model size (TrackAnythingEncodeKeyWithoutShrinkage) (float): 37.1 MB
- Number of parameters (TrackAnythingSegment): 30.1M
- Model size (TrackAnythingSegment) (float): 115 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| TrackAnythingEncodeKeyWithShrinkage | ONNX | float | Snapdragon® X Elite | 5.57 ms | 19 - 19 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 3.675 ms | 0 - 201 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | ONNX | float | Qualcomm® QCS8550 (Proxy) | 5.164 ms | 0 - 124 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | ONNX | float | Qualcomm® QCS9075 | 9.478 ms | 14 - 17 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.949 ms | 0 - 171 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.379 ms | 0 - 175 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | ONNX | float | Snapdragon® X2 Elite | 2.785 ms | 18 - 18 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 4.091 ms | 0 - 212 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 31.163 ms | 3 - 181 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 5.95 ms | 3 - 5 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Qualcomm® SA8775P | 9.582 ms | 3 - 181 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Qualcomm® QCS9075 | 11.198 ms | 2 - 29 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 11.71 ms | 3 - 197 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Qualcomm® SA7255P | 31.163 ms | 3 - 181 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Qualcomm® SA8295P | 10.413 ms | 3 - 165 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.282 ms | 0 - 182 MB | NPU |
| TrackAnythingEncodeKeyWithShrinkage | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.439 ms | 0 - 191 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | ONNX | float | Snapdragon® X Elite | 6.953 ms | 66 - 66 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 5.048 ms | 63 - 273 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | ONNX | float | Qualcomm® QCS8550 (Proxy) | 6.813 ms | 38 - 57 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | ONNX | float | Qualcomm® QCS9075 | 11.456 ms | 61 - 64 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.366 ms | 62 - 261 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.505 ms | 0 - 215 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | ONNX | float | Snapdragon® X2 Elite | 3.963 ms | 65 - 65 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 5.315 ms | 19 - 243 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 130.51 ms | 20 - 39 MB | GPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 7.598 ms | 20 - 23 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Qualcomm® SA8775P | 45.382 ms | 20 - 205 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Qualcomm® QCS9075 | 12.76 ms | 20 - 63 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 14.022 ms | 20 - 223 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Qualcomm® SA7255P | 130.51 ms | 20 - 39 MB | GPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Qualcomm® SA8295P | 12.716 ms | 20 - 185 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.406 ms | 19 - 210 MB | NPU |
| TrackAnythingEncodeKeyWithoutShrinkage | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.561 ms | 0 - 227 MB | NPU |
| TrackAnythingEncodeValue | ONNX | float | Snapdragon® X Elite | 12.612 ms | 38 - 38 MB | NPU |
| TrackAnythingEncodeValue | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 9.949 ms | 0 - 204 MB | NPU |
| TrackAnythingEncodeValue | ONNX | float | Qualcomm® QCS8550 (Proxy) | 15.647 ms | 0 - 50 MB | NPU |
| TrackAnythingEncodeValue | ONNX | float | Qualcomm® QCS9075 | 19.939 ms | 6 - 14 MB | NPU |
| TrackAnythingEncodeValue | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 8.417 ms | 9 - 181 MB | NPU |
| TrackAnythingEncodeValue | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 6.605 ms | 12 - 179 MB | NPU |
| TrackAnythingEncodeValue | ONNX | float | Snapdragon® X2 Elite | 6.053 ms | 39 - 39 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 8.3 ms | 2 - 291 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 51.37 ms | 3 - 223 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 11.833 ms | 3 - 5 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Qualcomm® SA8775P | 17.447 ms | 3 - 219 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Qualcomm® QCS9075 | 18.968 ms | 0 - 55 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 22.542 ms | 3 - 285 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Qualcomm® SA7255P | 51.37 ms | 3 - 223 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Qualcomm® SA8295P | 19.782 ms | 3 - 233 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 7.157 ms | 1 - 203 MB | NPU |
| TrackAnythingEncodeValue | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 6.263 ms | 2 - 210 MB | NPU |
| TrackAnythingSegment | ONNX | float | Snapdragon® X Elite | 23.52 ms | 37 - 37 MB | NPU |
| TrackAnythingSegment | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 17.504 ms | 26 - 272 MB | NPU |
| TrackAnythingSegment | ONNX | float | Qualcomm® QCS8550 (Proxy) | 23.034 ms | 0 - 43 MB | NPU |
| TrackAnythingSegment | ONNX | float | Qualcomm® QCS9075 | 37.018 ms | 21 - 45 MB | NPU |
| TrackAnythingSegment | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 13.654 ms | 7 - 223 MB | NPU |
| TrackAnythingSegment | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 10.478 ms | 10 - 223 MB | NPU |
| TrackAnythingSegment | ONNX | float | Snapdragon® X2 Elite | 11.859 ms | 38 - 38 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 17.596 ms | 0 - 286 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 103.823 ms | 2 - 198 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 23.346 ms | 2 - 4 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Qualcomm® SA8775P | 33.686 ms | 2 - 201 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Qualcomm® QCS9075 | 39.234 ms | 0 - 83 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 54.139 ms | 2 - 285 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Qualcomm® SA7255P | 103.823 ms | 2 - 198 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Qualcomm® SA8295P | 38.752 ms | 2 - 210 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 13.314 ms | 1 - 209 MB | NPU |
| TrackAnythingSegment | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 10.801 ms | 2 - 228 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.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support