Hugging Face's logo

qualcomm
/
YOLOv10-Detection

Object Detection
PyTorch
real_time
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
2beb8a6
·
verified ·
1 Parent(s): d64b39b

v0.48.0

Browse files

See https://github.com/qualcomm/ai-hub-models/releases/v0.48.0 for changelog.

Files changed (1) hide show
  1. README.md +64 -64
README.md CHANGED
@@ -15,18 +15,18 @@ pipeline_tag: object-detection
15
  Ultralytics YOLOv10 is a machine learning model that predicts bounding boxes and classes of objects in an image.
16
 
17
  This is based on the implementation of YOLOv10-Detection found [here](https://github.com/ultralytics/ultralytics/tree/main/ultralytics/models/yolo/detect).
18
- This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/yolov10_det) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
19
 
20
  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.
21
 
22
  ## Getting Started
23
  Due to licensing restrictions, we cannot distribute pre-exported model assets for this model.
24
- Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/yolov10_det) Python library to compile and export the model with your own:
25
  - Custom weights (e.g., fine-tuned checkpoints)
26
  - Custom input shapes
27
  - Target device and runtime configurations
28
 
29
- See our repository for [YOLOv10-Detection on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/yolov10_det) for usage instructions.
30
 
31
 
32
  ## Model Details
@@ -44,67 +44,67 @@ See our repository for [YOLOv10-Detection on GitHub](https://github.com/quic/ai-
44
  ## Performance Summary
45
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
46
  |---|---|---|---|---|---|---
47
- | YOLOv10-Detection | ONNX | float | Snapdragon® X Elite | 6.66 ms | 5 - 5 MB | NPU
48
- | YOLOv10-Detection | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 3.84 ms | 0 - 271 MB | NPU
49
- | YOLOv10-Detection | ONNX | float | Qualcomm® QCS8550 (Proxy) | 6.151 ms | 5 - 13 MB | NPU
50
- | YOLOv10-Detection | ONNX | float | Qualcomm® QCS9075 | 6.993 ms | 5 - 7 MB | NPU
51
- | YOLOv10-Detection | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.055 ms | 0 - 228 MB | NPU
52
- | YOLOv10-Detection | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.728 ms | 0 - 220 MB | NPU
53
- | YOLOv10-Detection | ONNX | float | Snapdragon® X2 Elite | 3.088 ms | 3 - 3 MB | NPU
54
- | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® X Elite | 5.521 ms | 2 - 2 MB | NPU
55
- | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.129 ms | 0 - 102 MB | NPU
56
- | YOLOv10-Detection | ONNX | w8a16 | Qualcomm® QCS6490 | 339.674 ms | 66 - 71 MB | CPU
57
- | YOLOv10-Detection | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 5.045 ms | 2 - 6 MB | NPU
58
- | YOLOv10-Detection | ONNX | w8a16 | Qualcomm® QCS9075 | 6.207 ms | 2 - 5 MB | NPU
59
- | YOLOv10-Detection | ONNX | w8a16 | Qualcomm® QCM6690 | 169.107 ms | 57 - 67 MB | CPU
60
- | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.293 ms | 0 - 77 MB | NPU
61
- | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 150.025 ms | 68 - 77 MB | CPU
62
- | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.085 ms | 0 - 79 MB | NPU
63
- | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® X2 Elite | 2.231 ms | 0 - 0 MB | NPU
64
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® X Elite | 14.185 ms | 8 - 8 MB | NPU
65
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Gen 3 Mobile | 8.431 ms | 8 - 269 MB | NPU
66
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Qualcomm® QCS6490 | 134.234 ms | 62 - 72 MB | CPU
67
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Qualcomm® QCS8550 (Proxy) | 11.985 ms | 6 - 17 MB | NPU
68
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Qualcomm® QCS9075 | 16.781 ms | 8 - 11 MB | NPU
69
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Qualcomm® QCM6690 | 79.415 ms | 52 - 63 MB | CPU
70
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Elite For Galaxy Mobile | 6.693 ms | 6 - 216 MB | NPU
71
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® 7 Gen 4 Mobile | 70.606 ms | 63 - 74 MB | CPU
72
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Elite Gen 5 Mobile | 6.157 ms | 9 - 230 MB | NPU
73
- | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® X2 Elite | 6.524 ms | 10 - 10 MB | NPU
74
- | YOLOv10-Detection | QNN_DLC | float | Snapdragon® X Elite | 4.389 ms | 5 - 5 MB | NPU
75
- | YOLOv10-Detection | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 2.939 ms | 5 - 216 MB | NPU
76
- | YOLOv10-Detection | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 13.093 ms | 0 - 182 MB | NPU
77
- | YOLOv10-Detection | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 4.001 ms | 2 - 6 MB | NPU
78
- | YOLOv10-Detection | QNN_DLC | float | Qualcomm® SA8775P | 5.44 ms | 0 - 183 MB | NPU
79
- | YOLOv10-Detection | QNN_DLC | float | Qualcomm® QCS9075 | 5.126 ms | 5 - 11 MB | NPU
80
- | YOLOv10-Detection | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 9.15 ms | 5 - 198 MB | NPU
81
- | YOLOv10-Detection | QNN_DLC | float | Qualcomm® SA7255P | 13.093 ms | 0 - 182 MB | NPU
82
- | YOLOv10-Detection | QNN_DLC | float | Qualcomm® SA8295P | 8.878 ms | 0 - 167 MB | NPU
83
- | YOLOv10-Detection | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.199 ms | 0 - 185 MB | NPU
84
- | YOLOv10-Detection | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.871 ms | 1 - 186 MB | NPU
85
- | YOLOv10-Detection | QNN_DLC | float | Snapdragon® X2 Elite | 2.604 ms | 5 - 5 MB | NPU
86
- | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® X Elite | 4.683 ms | 2 - 2 MB | NPU
87
- | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.007 ms | 0 - 200 MB | NPU
88
- | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 8.072 ms | 1 - 173 MB | NPU
89
- | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 4.296 ms | 2 - 20 MB | NPU
90
- | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® SA8775P | 4.926 ms | 0 - 176 MB | NPU
91
- | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 4.685 ms | 3 - 7 MB | NPU
92
- | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 18.646 ms | 2 - 182 MB | NPU
93
- | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® SA7255P | 8.072 ms | 1 - 173 MB | NPU
94
- | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.089 ms | 2 - 179 MB | NPU
95
- | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 4.441 ms | 0 - 178 MB | NPU
96
- | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 1.768 ms | 2 - 184 MB | NPU
97
- | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 2.202 ms | 2 - 2 MB | NPU
98
- | YOLOv10-Detection | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2.645 ms | 0 - 113 MB | NPU
99
- | YOLOv10-Detection | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 12.567 ms | 0 - 81 MB | NPU
100
- | YOLOv10-Detection | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 3.582 ms | 0 - 3 MB | NPU
101
- | YOLOv10-Detection | TFLITE | float | Qualcomm® SA8775P | 5.081 ms | 0 - 85 MB | NPU
102
- | YOLOv10-Detection | TFLITE | float | Qualcomm® QCS9075 | 4.824 ms | 0 - 13 MB | NPU
103
- | YOLOv10-Detection | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 7.769 ms | 0 - 197 MB | NPU
104
- | YOLOv10-Detection | TFLITE | float | Qualcomm® SA7255P | 12.567 ms | 0 - 81 MB | NPU
105
- | YOLOv10-Detection | TFLITE | float | Qualcomm® SA8295P | 8.141 ms | 0 - 171 MB | NPU
106
- | YOLOv10-Detection | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.984 ms | 0 - 94 MB | NPU
107
- | YOLOv10-Detection | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.582 ms | 0 - 91 MB | NPU
108
 
109
  ## License
110
  * The license for the original implementation of YOLOv10-Detection can be found
 
15
  Ultralytics YOLOv10 is a machine learning model that predicts bounding boxes and classes of objects in an image.
16
 
17
  This is based on the implementation of YOLOv10-Detection found [here](https://github.com/ultralytics/ultralytics/tree/main/ultralytics/models/yolo/detect).
18
+ 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/main/qai_hub_models/models/yolov10_det) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
19
 
20
  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.
21
 
22
  ## Getting Started
23
  Due to licensing restrictions, we cannot distribute pre-exported model assets for this model.
24
+ Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/yolov10_det) Python library to compile and export the model with your own:
25
  - Custom weights (e.g., fine-tuned checkpoints)
26
  - Custom input shapes
27
  - Target device and runtime configurations
28
 
29
+ See our repository for [YOLOv10-Detection on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/yolov10_det) for usage instructions.
30
 
31
 
32
  ## Model Details
 
44
  ## Performance Summary
45
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
46
  |---|---|---|---|---|---|---
47
+ | YOLOv10-Detection | ONNX | float | Snapdragon® X2 Elite | 3.099 ms | 3 - 3 MB | NPU
48
+ | YOLOv10-Detection | ONNX | float | Snapdragon® X Elite | 6.725 ms | 5 - 5 MB | NPU
49
+ | YOLOv10-Detection | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 3.833 ms | 2 - 277 MB | NPU
50
+ | YOLOv10-Detection | ONNX | float | Qualcomm® QCS8550 (Proxy) | 6.112 ms | 5 - 10 MB | NPU
51
+ | YOLOv10-Detection | ONNX | float | Qualcomm® QCS9075 | 7.008 ms | 5 - 7 MB | NPU
52
+ | YOLOv10-Detection | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.062 ms | 0 - 224 MB | NPU
53
+ | YOLOv10-Detection | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.719 ms | 3 - 223 MB | NPU
54
+ | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® X2 Elite | 2.234 ms | 0 - 0 MB | NPU
55
+ | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® X Elite | 5.547 ms | 2 - 2 MB | NPU
56
+ | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.16 ms | 0 - 103 MB | NPU
57
+ | YOLOv10-Detection | ONNX | w8a16 | Qualcomm® QCS6490 | 332.433 ms | 66 - 72 MB | CPU
58
+ | YOLOv10-Detection | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 5.056 ms | 2 - 7 MB | NPU
59
+ | YOLOv10-Detection | ONNX | w8a16 | Qualcomm® QCS9075 | 6.102 ms | 2 - 5 MB | NPU
60
+ | YOLOv10-Detection | ONNX | w8a16 | Qualcomm® QCM6690 | 168.326 ms | 57 - 65 MB | CPU
61
+ | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.301 ms | 0 - 72 MB | NPU
62
+ | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 149.606 ms | 67 - 76 MB | CPU
63
+ | YOLOv10-Detection | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.086 ms | 0 - 79 MB | NPU
64
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® X2 Elite | 6.753 ms | 10 - 10 MB | NPU
65
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® X Elite | 14.849 ms | 8 - 8 MB | NPU
66
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Gen 3 Mobile | 8.563 ms | 9 - 234 MB | NPU
67
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Qualcomm® QCS6490 | 140.766 ms | 63 - 73 MB | CPU
68
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Qualcomm® QCS8550 (Proxy) | 11.809 ms | 7 - 15 MB | NPU
69
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Qualcomm® QCS9075 | 16.39 ms | 8 - 10 MB | NPU
70
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Qualcomm® QCM6690 | 80.941 ms | 53 - 64 MB | CPU
71
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Elite For Galaxy Mobile | 6.459 ms | 6 - 196 MB | NPU
72
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® 7 Gen 4 Mobile | 71.711 ms | 63 - 75 MB | CPU
73
+ | YOLOv10-Detection | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Elite Gen 5 Mobile | 6.341 ms | 9 - 323 MB | NPU
74
+ | YOLOv10-Detection | QNN_DLC | float | Snapdragon® X2 Elite | 2.591 ms | 5 - 5 MB | NPU
75
+ | YOLOv10-Detection | QNN_DLC | float | Snapdragon® X Elite | 4.388 ms | 5 - 5 MB | NPU
76
+ | YOLOv10-Detection | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 2.943 ms | 5 - 217 MB | NPU
77
+ | YOLOv10-Detection | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 13.081 ms | 1 - 183 MB | NPU
78
+ | YOLOv10-Detection | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 3.98 ms | 5 - 6 MB | NPU
79
+ | YOLOv10-Detection | QNN_DLC | float | Qualcomm® SA8775P | 5.441 ms | 0 - 184 MB | NPU
80
+ | YOLOv10-Detection | QNN_DLC | float | Qualcomm® QCS9075 | 5.121 ms | 5 - 11 MB | NPU
81
+ | YOLOv10-Detection | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 9.208 ms | 4 - 198 MB | NPU
82
+ | YOLOv10-Detection | QNN_DLC | float | Qualcomm® SA7255P | 13.081 ms | 1 - 183 MB | NPU
83
+ | YOLOv10-Detection | QNN_DLC | float | Qualcomm® SA8295P | 8.855 ms | 1 - 168 MB | NPU
84
+ | YOLOv10-Detection | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.208 ms | 0 - 185 MB | NPU
85
+ | YOLOv10-Detection | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.872 ms | 5 - 189 MB | NPU
86
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 2.155 ms | 2 - 2 MB | NPU
87
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® X Elite | 4.674 ms | 2 - 2 MB | NPU
88
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.003 ms | 2 - 204 MB | NPU
89
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 8.162 ms | 1 - 173 MB | NPU
90
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 4.289 ms | 2 - 4 MB | NPU
91
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® SA8775P | 4.914 ms | 0 - 176 MB | NPU
92
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 4.83 ms | 0 - 4 MB | NPU
93
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 19.745 ms | 2 - 182 MB | NPU
94
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Qualcomm® SA7255P | 8.162 ms | 1 - 173 MB | NPU
95
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.131 ms | 0 - 175 MB | NPU
96
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 4.399 ms | 2 - 177 MB | NPU
97
+ | YOLOv10-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 1.798 ms | 2 - 184 MB | NPU
98
+ | YOLOv10-Detection | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2.658 ms | 0 - 114 MB | NPU
99
+ | YOLOv10-Detection | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 12.566 ms | 0 - 81 MB | NPU
100
+ | YOLOv10-Detection | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 3.564 ms | 0 - 3 MB | NPU
101
+ | YOLOv10-Detection | TFLITE | float | Qualcomm® SA8775P | 5.106 ms | 0 - 84 MB | NPU
102
+ | YOLOv10-Detection | TFLITE | float | Qualcomm® QCS9075 | 4.816 ms | 0 - 13 MB | NPU
103
+ | YOLOv10-Detection | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 7.702 ms | 0 - 196 MB | NPU
104
+ | YOLOv10-Detection | TFLITE | float | Qualcomm® SA7255P | 12.566 ms | 0 - 81 MB | NPU
105
+ | YOLOv10-Detection | TFLITE | float | Qualcomm® SA8295P | 8.133 ms | 0 - 170 MB | NPU
106
+ | YOLOv10-Detection | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.991 ms | 0 - 84 MB | NPU
107
+ | YOLOv10-Detection | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.58 ms | 0 - 92 MB | NPU
108
 
109
  ## License
110
  * The license for the original implementation of YOLOv10-Detection can be found