Fast Segment Anything
Paper • 2306.12156 • Published • 36
FastSam-S: Optimized for Qualcomm Devices
The Fast Segment Anything Model (FastSAM) is a novel, real-time CNN-based solution for the Segment Anything task. This task is designed to segment any object within an image based on various possible user interaction prompts. The model performs competitively despite significantly reduced computation, making it a practical choice for a variety of vision tasks.
This is based on the implementation of FastSam-S 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 |
| QNN_DLC | float | Universal | QAIRT 2.43 | Download |
| TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | Download |
For more device-specific assets and performance metrics, visit FastSam-S 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 FastSam-S on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.semantic_segmentation
Model Stats:
- Model checkpoint: fastsam-s.pt
- Inference latency: RealTime
- Input resolution: 640x640
- Number of parameters: 11.8M
- Model size (float): 45.1 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| FastSam-S | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.367 ms | 1 - 211 MB | NPU |
| FastSam-S | ONNX | float | Snapdragon® X2 Elite | 4.37 ms | 20 - 20 MB | NPU |
| FastSam-S | ONNX | float | Snapdragon® X Elite | 8.625 ms | 19 - 19 MB | NPU |
| FastSam-S | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 5.93 ms | 15 - 285 MB | NPU |
| FastSam-S | ONNX | float | Qualcomm® QCS8550 (Proxy) | 8.037 ms | 0 - 71 MB | NPU |
| FastSam-S | ONNX | float | Qualcomm® QCS9075 | 13.122 ms | 11 - 14 MB | NPU |
| FastSam-S | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.514 ms | 10 - 235 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.114 ms | 4 - 192 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® X2 Elite | 4.437 ms | 5 - 5 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® X Elite | 7.964 ms | 5 - 5 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 5.524 ms | 5 - 215 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 38.665 ms | 0 - 181 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 7.434 ms | 5 - 9 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® SA8775P | 11.279 ms | 0 - 189 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS9075 | 10.884 ms | 5 - 15 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 15.021 ms | 5 - 212 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® SA7255P | 38.665 ms | 0 - 181 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® SA8295P | 13.896 ms | 0 - 176 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.297 ms | 5 - 190 MB | NPU |
| FastSam-S | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.887 ms | 0 - 201 MB | NPU |
| FastSam-S | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 5.117 ms | 3 - 122 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 37.716 ms | 4 - 81 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 6.817 ms | 1 - 5 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® SA8775P | 10.637 ms | 4 - 85 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS9075 | 10.598 ms | 4 - 39 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 13.972 ms | 2 - 230 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® SA7255P | 37.716 ms | 4 - 81 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® SA8295P | 13.104 ms | 4 - 200 MB | NPU |
| FastSam-S | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.855 ms | 0 - 93 MB | NPU |
License
- The license for the original implementation of FastSam-S 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.