Hugging Face's logo

qualcomm
/
Sequencer2D

Image Classification
PyTorch
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
4bcabbc
·
verified ·
1 Parent(s): 3d1fc46

v0.58.0

Browse files

See https://github.com/qualcomm/ai-hub-models/releases/v0.58.0 for changelog.

Files changed (2) hide show
  1. README.md +60 -29
  2. release_assets.json +4 -4
README.md CHANGED
@@ -14,7 +14,7 @@ pipeline_tag: image-classification
14
  Sequencer2D is a vision transformer model that can classify images from the Imagenet dataset.
15
 
16
  This is based on the implementation of Sequencer2D found [here](https://github.com/okojoalg/sequencer).
17
- 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/v0.57.3/src/qai_hub_models/models/sequencer2d) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
18
 
19
  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.
20
 
@@ -27,23 +27,23 @@ Below are pre-exported model assets ready for deployment.
27
 
28
  | Runtime | Precision | Chipset | SDK Versions | Download |
29
  |---|---|---|---|---|
30
- | ONNX | float | Universal | QAIRT 2.45, ONNX Runtime 1.25.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.57.3/sequencer2d-onnx-float.zip)
31
- | QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.57.3/sequencer2d-qnn_dlc-float.zip)
32
- | TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.57.3/sequencer2d-tflite-float.zip)
33
 
34
  For more device-specific assets and performance metrics, visit **[Sequencer2D on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/sequencer2d)**.
35
 
36
 
37
  ### Option 2: Export with Custom Configurations
38
 
39
- Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.57.3/src/qai_hub_models/models/sequencer2d) Python library to compile and export the model with your own:
40
  - Custom weights (e.g., fine-tuned checkpoints)
41
  - Custom input shapes
42
  - Target device and runtime configurations
43
 
44
  This option is ideal if you need to customize the model beyond the default configuration provided here.
45
 
46
- See our repository for [Sequencer2D on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.57.3/src/qai_hub_models/models/sequencer2d) for usage instructions.
47
 
48
  ## Model Details
49
 
@@ -59,31 +59,62 @@ See our repository for [Sequencer2D on GitHub](https://github.com/qualcomm/ai-hu
59
  ## Performance Summary
60
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
61
  |---|---|---|---|---|---|---
62
- | Sequencer2D | ONNX | float | Snapdragon® X2 Elite | 9.841 ms | 182 - 182 MB | NPU
63
- | Sequencer2D | ONNX | float | Snapdragon® X Elite | 16.903 ms | 174 - 174 MB | NPU
64
- | Sequencer2D | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 11.197 ms | 1 - 950 MB | NPU
65
- | Sequencer2D | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 22.62 ms | 1 - 650 MB | NPU
66
- | Sequencer2D | ONNX | float | Qualcomm® QCS8550 (Proxy) | 16.478 ms | 0 - 72 MB | NPU
67
- | Sequencer2D | ONNX | float | Qualcomm® QCS8450 | 22.62 ms | 1 - 650 MB | NPU
68
- | Sequencer2D | ONNX | float | Snapdragon® 8 Elite Mobile | 9.593 ms | 1 - 739 MB | NPU
69
- | Sequencer2D | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 8.029 ms | 1 - 891 MB | NPU
70
- | Sequencer2D | ONNX | float | Qualcomm® QCS9075 | 19.389 ms | 1 - 45 MB | NPU
71
- | Sequencer2D | ONNX | float | Qualcomm® QCS8750 | 9.593 ms | 1 - 739 MB | NPU
72
- | Sequencer2D | ONNX | float | Qualcomm® QCS7181 | 16.903 ms | 174 - 174 MB | NPU
73
- | Sequencer2D | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 12.104 ms | 0 - 988 MB | NPU
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
74
  | Sequencer2D | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 21.228 ms | 0 - 732 MB | NPU
75
- | Sequencer2D | TFLITE | float | Qualcomm® QCS8275 | 37.429 ms | 0 - 795 MB | NPU
76
- | Sequencer2D | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 18.121 ms | 0 - 8 MB | NPU
77
- | Sequencer2D | TFLITE | float | Qualcomm® SA8775P | 194.134 ms | 6 - 58 MB | CPU
78
- | Sequencer2D | TFLITE | float | Qualcomm® SA8650P | 194.134 ms | 6 - 58 MB | CPU
79
- | Sequencer2D | TFLITE | float | Qualcomm® SA8255P | 194.134 ms | 6 - 58 MB | CPU
80
  | Sequencer2D | TFLITE | float | Qualcomm® QCS8450 | 21.228 ms | 0 - 732 MB | NPU
81
- | Sequencer2D | TFLITE | float | Snapdragon® 8 Elite Mobile | 9.358 ms | 0 - 794 MB | NPU
82
- | Sequencer2D | TFLITE | float | Qualcomm® SA8295P | 20.517 ms | 0 - 486 MB | NPU
83
- | Sequencer2D | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 7.232 ms | 0 - 761 MB | NPU
84
- | Sequencer2D | TFLITE | float | Qualcomm® SA7255P | 37.429 ms | 0 - 795 MB | NPU
85
- | Sequencer2D | TFLITE | float | Qualcomm® QCS9075 | 20.11 ms | 0 - 75 MB | NPU
86
- | Sequencer2D | TFLITE | float | Qualcomm® QCS8750 | 9.358 ms | 0 - 794 MB | NPU
 
 
 
 
 
 
 
 
 
 
 
 
 
 
87
 
88
  ## License
89
  * The license for the original implementation of Sequencer2D can be found
 
14
  Sequencer2D is a vision transformer model that can classify images from the Imagenet dataset.
15
 
16
  This is based on the implementation of Sequencer2D found [here](https://github.com/okojoalg/sequencer).
17
+ 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/v0.58.0/src/qai_hub_models/models/sequencer2d) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
18
 
19
  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.
20
 
 
27
 
28
  | Runtime | Precision | Chipset | SDK Versions | Download |
29
  |---|---|---|---|---|
30
+ | ONNX | float | Universal | QAIRT 2.45, ONNX Runtime 1.25.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.58.0/sequencer2d-onnx-float.zip)
31
+ | QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.58.0/sequencer2d-qnn_dlc-float.zip)
32
+ | TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.58.0/sequencer2d-tflite-float.zip)
33
 
34
  For more device-specific assets and performance metrics, visit **[Sequencer2D on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/sequencer2d)**.
35
 
36
 
37
  ### Option 2: Export with Custom Configurations
38
 
39
+ Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.58.0/src/qai_hub_models/models/sequencer2d) Python library to compile and export the model with your own:
40
  - Custom weights (e.g., fine-tuned checkpoints)
41
  - Custom input shapes
42
  - Target device and runtime configurations
43
 
44
  This option is ideal if you need to customize the model beyond the default configuration provided here.
45
 
46
+ See our repository for [Sequencer2D on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.58.0/src/qai_hub_models/models/sequencer2d) for usage instructions.
47
 
48
  ## Model Details
49
 
 
59
  ## Performance Summary
60
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
61
  |---|---|---|---|---|---|---
62
+ | Sequencer2D | ONNX | float | Snapdragon® X2 Elite | 9.818 ms | 2 - 2 MB | NPU
63
+ | Sequencer2D | ONNX | float | Snapdragon® X Elite | 16.963 ms | 60 - 60 MB | NPU
64
+ | Sequencer2D | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 10.918 ms | 0 - 961 MB | NPU
65
+ | Sequencer2D | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 22.469 ms | 0 - 659 MB | NPU
66
+ | Sequencer2D | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 15.866 ms | 0 - 73 MB | NPU
67
+ | Sequencer2D | ONNX | float | Qualcomm® QCS8450 | 22.469 ms | 0 - 659 MB | NPU
68
+ | Sequencer2D | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 8.049 ms | 0 - 893 MB | NPU
69
+ | Sequencer2D | ONNX | float | Snapdragon® 8 Elite Mobile | 9.539 ms | 1 - 739 MB | NPU
70
+ | Sequencer2D | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 24.869 ms | 0 - 4 MB | NPU
71
+ | Sequencer2D | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 9.539 ms | 1 - 739 MB | NPU
72
+ | Sequencer2D | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 16.963 ms | 60 - 60 MB | NPU
73
+ | Sequencer2D | QNN_DLC | float | Snapdragon® X2 Elite | 11.815 ms | 1 - 1 MB | NPU
74
+ | Sequencer2D | QNN_DLC | float | Snapdragon® X Elite | 21.496 ms | 1 - 1 MB | NPU
75
+ | Sequencer2D | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 14.115 ms | 0 - 2192 MB | NPU
76
+ | Sequencer2D | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 25.546 ms | 0 - 798 MB | NPU
77
+ | Sequencer2D | QNN_DLC | float | Qualcomm® QCS8275 | 51.741 ms | 1 - 996 MB | NPU
78
+ | Sequencer2D | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 20.399 ms | 1 - 4 MB | NPU
79
+ | Sequencer2D | QNN_DLC | float | Qualcomm® SA8775P | 24.1 ms | 1 - 909 MB | NPU
80
+ | Sequencer2D | QNN_DLC | float | Qualcomm® SA8650P | 24.1 ms | 1 - 909 MB | NPU
81
+ | Sequencer2D | QNN_DLC | float | Qualcomm® SA8255P | 24.1 ms | 1 - 909 MB | NPU
82
+ | Sequencer2D | QNN_DLC | float | Qualcomm® QCS8450 | 25.546 ms | 0 - 798 MB | NPU
83
+ | Sequencer2D | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 9.18 ms | 0 - 867 MB | NPU
84
+ | Sequencer2D | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 12.127 ms | 1 - 1035 MB | NPU
85
+ | Sequencer2D | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 23.759 ms | 1 - 3 MB | NPU
86
+ | Sequencer2D | QNN_DLC | float | Qualcomm® SA7255P | 51.741 ms | 1 - 996 MB | NPU
87
+ | Sequencer2D | QNN_DLC | float | Qualcomm® SA8295P | 26.032 ms | 1 - 666 MB | NPU
88
+ | Sequencer2D | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 12.127 ms | 1 - 1035 MB | NPU
89
+ | Sequencer2D | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 21.496 ms | 1 - 1 MB | NPU
90
+ | Sequencer2D | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 11.981 ms | 119 - 1098 MB | NPU
91
  | Sequencer2D | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 21.228 ms | 0 - 732 MB | NPU
92
+ | Sequencer2D | TFLITE | float | Qualcomm® QCS8275 | 37.333 ms | 0 - 795 MB | NPU
93
+ | Sequencer2D | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 16.629 ms | 0 - 9 MB | NPU
94
+ | Sequencer2D | TFLITE | float | Qualcomm® SA8775P | 19.071 ms | 0 - 795 MB | NPU
95
+ | Sequencer2D | TFLITE | float | Qualcomm® SA8650P | 19.071 ms | 0 - 795 MB | NPU
96
+ | Sequencer2D | TFLITE | float | Qualcomm® SA8255P | 19.071 ms | 0 - 795 MB | NPU
97
  | Sequencer2D | TFLITE | float | Qualcomm® QCS8450 | 21.228 ms | 0 - 732 MB | NPU
98
+ | Sequencer2D | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 7.234 ms | 0 - 763 MB | NPU
99
+ | Sequencer2D | TFLITE | float | Snapdragon® 8 Elite Mobile | 9.291 ms | 0 - 787 MB | NPU
100
+ | Sequencer2D | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 18.257 ms | 0 - 76 MB | NPU
101
+ | Sequencer2D | TFLITE | float | Qualcomm® SA7255P | 37.333 ms | 0 - 795 MB | NPU
102
+ | Sequencer2D | TFLITE | float | Qualcomm® SA8295P | 20.544 ms | 0 - 485 MB | NPU
103
+ | Sequencer2D | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 9.291 ms | 0 - 787 MB | NPU
104
+ | Sequencer2D | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 13.509 ms | 0 - 947 MB | NPU
105
+ | Sequencer2D | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 21.837 ms | 0 - 814 MB | NPU
106
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® QCS8275 | 36.319 ms | 0 - 862 MB | NPU
107
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 19.077 ms | 0 - 11 MB | NPU
108
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® SA8775P | 20.358 ms | 0 - 862 MB | NPU
109
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® SA8650P | 20.358 ms | 0 - 862 MB | NPU
110
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® SA8255P | 20.358 ms | 0 - 862 MB | NPU
111
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® QCS8450 | 21.837 ms | 0 - 814 MB | NPU
112
+ | Sequencer2D | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 10.989 ms | 0 - 802 MB | NPU
113
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 19.774 ms | 0 - 64 MB | NPU
114
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® SA7255P | 36.319 ms | 0 - 862 MB | NPU
115
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® SA8295P | 22.921 ms | 0 - 617 MB | NPU
116
+ | Sequencer2D | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 9.076 ms | 0 - 820 MB | NPU
117
+ | Sequencer2D | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 10.989 ms | 0 - 802 MB | NPU
118
 
119
  ## License
120
  * The license for the original implementation of Sequencer2D can be found
release_assets.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "version": "0.57.3",
3
  "precisions": {
4
  "float": {
5
  "universal_assets": {
@@ -8,20 +8,20 @@
8
  "qairt": "2.45.0.260326154327",
9
  "litert": "1.4.4"
10
  },
11
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.57.3/sequencer2d-tflite-float.zip"
12
  },
13
  "qnn_dlc": {
14
  "tool_versions": {
15
  "qairt": "2.45.0.260326154327"
16
  },
17
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.57.3/sequencer2d-qnn_dlc-float.zip"
18
  },
19
  "onnx": {
20
  "tool_versions": {
21
  "qairt": "2.45.0.260326154327",
22
  "onnx_runtime": "1.25.0"
23
  },
24
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.57.3/sequencer2d-onnx-float.zip"
25
  }
26
  }
27
  }
 
1
  {
2
+ "version": "0.58.0",
3
  "precisions": {
4
  "float": {
5
  "universal_assets": {
 
8
  "qairt": "2.45.0.260326154327",
9
  "litert": "1.4.4"
10
  },
11
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.58.0/sequencer2d-tflite-float.zip"
12
  },
13
  "qnn_dlc": {
14
  "tool_versions": {
15
  "qairt": "2.45.0.260326154327"
16
  },
17
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.58.0/sequencer2d-qnn_dlc-float.zip"
18
  },
19
  "onnx": {
20
  "tool_versions": {
21
  "qairt": "2.45.0.260326154327",
22
  "onnx_runtime": "1.25.0"
23
  },
24
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sequencer2d/releases/v0.58.0/sequencer2d-onnx-float.zip"
25
  }
26
  }
27
  }