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 1.25.0 | Download |
| QNN_DLC | float | Universal | QAIRT 2.45 | Download |
| TFLITE | float | Universal | QAIRT 2.45 | 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 | 3.543 ms | 1 - 208 MB | NPU |
| FastSam-S | ONNX | float | Snapdragon® X2 Elite | 4.522 ms | 207 - 207 MB | NPU |
| FastSam-S | ONNX | float | Snapdragon® X Elite | 8.536 ms | 176 - 176 MB | NPU |
| FastSam-S | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 6.052 ms | 17 - 250 MB | NPU |
| FastSam-S | ONNX | float | Qualcomm® QCS8550 (Proxy) | 8.249 ms | 0 - 27 MB | NPU |
| FastSam-S | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.714 ms | 13 - 211 MB | NPU |
| FastSam-S | ONNX | float | Qualcomm® QCS9075 | 13.529 ms | 16 - 61 MB | NPU |
| FastSam-S | ONNX | float | Qualcomm® QCS8750 | 4.714 ms | 13 - 211 MB | NPU |
| FastSam-S | ONNX | float | Qualcomm® QCS7181 | 8.536 ms | 176 - 176 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.151 ms | 1 - 194 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® X2 Elite | 4.586 ms | 5 - 5 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® X Elite | 8.249 ms | 5 - 5 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 5.847 ms | 4 - 217 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS8275 | 39.231 ms | 2 - 188 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 7.716 ms | 5 - 7 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® SA8775P | 11.615 ms | 1 - 180 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® SA8650P | 11.615 ms | 1 - 180 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® SA8255P | 11.615 ms | 1 - 180 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 16.083 ms | 4 - 213 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® SA7255P | 39.231 ms | 2 - 188 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® SA8295P | 13.794 ms | 0 - 178 MB | NPU |
| FastSam-S | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.486 ms | 0 - 190 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS9075 | 10.943 ms | 5 - 15 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS8750 | 4.486 ms | 0 - 190 MB | NPU |
| FastSam-S | QNN_DLC | float | Qualcomm® QCS7181 | 8.249 ms | 5 - 5 MB | NPU |
| FastSam-S | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.996 ms | 0 - 98 MB | NPU |
| FastSam-S | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 5.164 ms | 3 - 116 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS8275 | 37.802 ms | 4 - 86 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 6.948 ms | 4 - 9 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® SA8775P | 10.736 ms | 4 - 90 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® SA8650P | 10.736 ms | 4 - 90 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® SA8255P | 10.736 ms | 4 - 90 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 14.147 ms | 3 - 231 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® SA7255P | 37.802 ms | 4 - 86 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® SA8295P | 13.089 ms | 4 - 195 MB | NPU |
| FastSam-S | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.946 ms | 4 - 101 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS9075 | 10.677 ms | 4 - 39 MB | NPU |
| FastSam-S | TFLITE | float | Qualcomm® QCS8750 | 3.946 ms | 4 - 101 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.