Hugging Face's logo

qualcomm
/
FastSam-X

Image Segmentation
PyTorch
android
Model card Files
xet
 Community
FastSam-X
File size: 11,729 Bytes
01a0a41
 
5727858
01a0a41
 
fbb96d2
01a0a41
 
 
b504f57
01a0a41
f0b3ea8
673bd8a
01a0a41
 
4b65a8c
c125946
f0b3ea8
 
 
 
 
 
 
 
 
 
 
 
c125946
 
 
 
 
f0b3ea8
 
 
 
 
 
c125946
f0b3ea8
 
 
 
 
 
c125946
f0b3ea8
 
 
 
 
 
 
 
c125946
f0b3ea8
c125946
f0b3ea8
 
 
 
c125946
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
94d4192
01a0a41
396f835
4b65a8c
94d4192
01a0a41
 
4b65a8c
01a0a41
 
dc89166
01a0a41
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
---
library_name: pytorch
license: other
tags:
- android
pipeline_tag: image-segmentation

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/fastsam_x/web-assets/model_demo.png)

# FastSam-X: 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-X found [here](https://github.com/CASIA-LMC-Lab/FastSAM).
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/blob/v0.60.0/src/qai_hub_models/models/fastsam_x) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).

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.

## 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 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/fastsam_x/releases/v0.60.0/fastsam_x-onnx-float.zip)
| ONNX | w8a16 | Universal | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/fastsam_x/releases/v0.60.0/fastsam_x-onnx-w8a16.zip)
| QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/fastsam_x/releases/v0.60.0/fastsam_x-qnn_dlc-float.zip)
| QNN_DLC | w8a16 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/fastsam_x/releases/v0.60.0/fastsam_x-qnn_dlc-w8a16.zip)
| TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/fastsam_x/releases/v0.60.0/fastsam_x-tflite-float.zip)

For more device-specific assets and performance metrics, visit **[FastSam-X on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/fastsam_x)**.


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/fastsam_x) 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-X on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/fastsam_x) for usage instructions.

## Model Details

**Model Type:** Model_use_case.semantic_segmentation

**Model Stats:**
- Inference latency: RealTime
- Input resolution: 640x640
- Model checkpoint: fastsam-x.pt
- Model size (float): 276 MB
- Number of parameters: 72.2M

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| FastSam-X | ONNX | float | Snapdragon® X2 Elite | 24.581 ms | 17 - 17 MB | NPU
| FastSam-X | ONNX | float | Snapdragon® X Elite | 47.017 ms | 140 - 140 MB | NPU
| FastSam-X | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 36.479 ms | 1 - 319 MB | NPU
| FastSam-X | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 100.223 ms | 1 - 400 MB | NPU
| FastSam-X | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 84.659 ms | 5 - 13 MB | NPU
| FastSam-X | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 49.427 ms | 15 - 18 MB | NPU
| FastSam-X | ONNX | float | Qualcomm® QCS8450 | 100.223 ms | 1 - 400 MB | NPU
| FastSam-X | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 72.433 ms | 11 - 19 MB | NPU
| FastSam-X | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 47.017 ms | 140 - 140 MB | NPU
| FastSam-X | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 27.6 ms | 13 - 245 MB | NPU
| FastSam-X | ONNX | float | Snapdragon® 8 Elite Mobile | 27.6 ms | 13 - 245 MB | NPU
| FastSam-X | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 20.044 ms | 12 - 265 MB | NPU
| FastSam-X | ONNX | w8a16 | Snapdragon® X2 Elite | 15.553 ms | 9 - 9 MB | NPU
| FastSam-X | ONNX | w8a16 | Snapdragon® X Elite | 33.689 ms | 71 - 71 MB | NPU
| FastSam-X | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 24.645 ms | 0 - 452 MB | NPU
| FastSam-X | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 56.545 ms | 9 - 462 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 183.5 ms | 6 - 11 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 33.821 ms | 6 - 12 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 34.202 ms | 0 - 81 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® QCS8450 | 56.545 ms | 9 - 462 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 31.286 ms | 6 - 11 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 33.689 ms | 71 - 71 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 692.601 ms | 0 - 546 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 61.253 ms | 8 - 431 MB | NPU
| FastSam-X | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 19.867 ms | 6 - 364 MB | NPU
| FastSam-X | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 19.867 ms | 6 - 364 MB | NPU
| FastSam-X | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 12.921 ms | 6 - 370 MB | NPU
| FastSam-X | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 61.253 ms | 8 - 431 MB | NPU
| FastSam-X | QNN_DLC | float | Snapdragon® X2 Elite | 23.1 ms | 5 - 5 MB | NPU
| FastSam-X | QNN_DLC | float | Snapdragon® X Elite | 44.037 ms | 5 - 5 MB | NPU
| FastSam-X | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 32.798 ms | 5 - 303 MB | NPU
| FastSam-X | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 91.529 ms | 5 - 389 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 80.408 ms | 5 - 16 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 279.924 ms | 1 - 215 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 43.181 ms | 5 - 7 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® SA8775P | 68.471 ms | 1 - 213 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® SA8650P | 68.471 ms | 1 - 213 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® SA8255P | 68.471 ms | 1 - 213 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® QCS8450 | 91.529 ms | 5 - 389 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 68.265 ms | 5 - 15 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 44.037 ms | 5 - 5 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 25.252 ms | 0 - 221 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® SA7255P | 279.924 ms | 1 - 215 MB | NPU
| FastSam-X | QNN_DLC | float | Qualcomm® SA8295P | 77.684 ms | 0 - 298 MB | NPU
| FastSam-X | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 25.252 ms | 0 - 221 MB | NPU
| FastSam-X | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 17.981 ms | 1 - 233 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 15.164 ms | 2 - 2 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Snapdragon® X Elite | 31.387 ms | 2 - 2 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 22.319 ms | 2 - 419 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Snapdragon® 8 Gen 1 Mobile | 48.903 ms | 2 - 417 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 178.827 ms | 2 - 8 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 30.656 ms | 2 - 9 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 86.223 ms | 1 - 320 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 31.329 ms | 2 - 5 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® SA8775P | 28.445 ms | 0 - 322 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® SA8650P | 28.445 ms | 0 - 322 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® SA8255P | 28.445 ms | 0 - 322 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® QCS8450 | 48.903 ms | 2 - 417 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 31.742 ms | 1 - 7 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 31.387 ms | 2 - 2 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 655.926 ms | 3 - 481 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 53.476 ms | 2 - 398 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 18.238 ms | 2 - 332 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® SA7255P | 86.223 ms | 1 - 320 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Qualcomm® SA8295P | 45.07 ms | 0 - 319 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 18.238 ms | 2 - 332 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 12.218 ms | 2 - 333 MB | NPU
| FastSam-X | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 53.476 ms | 2 - 398 MB | NPU
| FastSam-X | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 32.55 ms | 3 - 417 MB | NPU
| FastSam-X | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 94.487 ms | 5 - 506 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 80.236 ms | 4 - 159 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 279.398 ms | 5 - 257 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 43.235 ms | 4 - 7 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® SA8775P | 67.958 ms | 5 - 258 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® SA8650P | 67.958 ms | 5 - 258 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® SA8255P | 67.958 ms | 5 - 258 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® QCS8450 | 94.487 ms | 5 - 506 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 67.971 ms | 4 - 158 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 25.727 ms | 4 - 263 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® SA7255P | 279.398 ms | 5 - 257 MB | NPU
| FastSam-X | TFLITE | float | Qualcomm® SA8295P | 77.149 ms | 5 - 335 MB | NPU
| FastSam-X | TFLITE | float | Snapdragon® 8 Elite Mobile | 25.727 ms | 4 - 263 MB | NPU
| FastSam-X | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 17.404 ms | 0 - 262 MB | NPU

## License
* The license for the original implementation of FastSam-X can be found
  [here](https://github.com/CASIA-LMC-Lab/FastSAM/blob/main/LICENSE).

## References
* [Fast Segment Anything](https://arxiv.org/abs/2306.12156)
* [Source Model Implementation](https://github.com/CASIA-LMC-Lab/FastSAM)

## Community
* Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
* For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).