Hugging Face's logo

qualcomm
/
Video-MAE

Video Classification
PyTorch
backbone
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
e541dd7
·
verified ·
1 Parent(s): 425f873

v0.49.1

Browse files

See https://github.com/qualcomm/ai-hub-models/releases/v0.49.1 for changelog.

Files changed (1) hide show
  1. README.md +45 -46
README.md CHANGED
@@ -15,7 +15,7 @@ pipeline_tag: video-classification
15
  Video MAE (Masked Auto Encoder) is a network for doing video classification that uses the ViT (Vision Transformer) backbone.
16
 
17
  This is based on the implementation of Video-MAE found [here](https://github.com/MCG-NJU/VideoMAE).
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/video_mae) 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
 
@@ -28,25 +28,25 @@ Below are pre-exported model assets ready for deployment.
28
 
29
  | Runtime | Precision | Chipset | SDK Versions | Download |
30
  |---|---|---|---|---|
31
- | ONNX | float | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.48.0/video_mae-onnx-float.zip)
32
- | ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.48.0/video_mae-onnx-w8a16.zip)
33
- | QNN_DLC | float | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.48.0/video_mae-qnn_dlc-float.zip)
34
- | QNN_DLC | w8a16 | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.48.0/video_mae-qnn_dlc-w8a16.zip)
35
- | TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.48.0/video_mae-tflite-float.zip)
36
 
37
  For more device-specific assets and performance metrics, visit **[Video-MAE on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/video_mae)**.
38
 
39
 
40
  ### Option 2: Export with Custom Configurations
41
 
42
- Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/video_mae) Python library to compile and export the model with your own:
43
  - Custom weights (e.g., fine-tuned checkpoints)
44
  - Custom input shapes
45
  - Target device and runtime configurations
46
 
47
  This option is ideal if you need to customize the model beyond the default configuration provided here.
48
 
49
- See our repository for [Video-MAE on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/video_mae) for usage instructions.
50
 
51
  ## Model Details
52
 
@@ -61,44 +61,43 @@ See our repository for [Video-MAE on GitHub](https://github.com/qualcomm/ai-hub-
61
  ## Performance Summary
62
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
63
  |---|---|---|---|---|---|---
64
- | Video-MAE | ONNX | float | Snapdragon® X2 Elite | 465.083 ms | 188 - 188 MB | NPU
65
- | Video-MAE | ONNX | float | Snapdragon® X Elite | 668.776 ms | 188 - 188 MB | NPU
66
- | Video-MAE | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 388.833 ms | 9 - 1257 MB | NPU
67
- | Video-MAE | ONNX | float | Qualcomm® QCS8550 (Proxy) | 644.907 ms | 0 - 219 MB | NPU
68
- | Video-MAE | ONNX | float | Qualcomm® QCS9075 | 686.387 ms | 9 - 21 MB | NPU
69
- | Video-MAE | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 394.489 ms | 0 - 979 MB | NPU
70
- | Video-MAE | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 444.692 ms | 9 - 1065 MB | NPU
71
- | Video-MAE | ONNX | w8a16 | Snapdragon® X2 Elite | 323.006 ms | 123 - 123 MB | NPU
72
- | Video-MAE | ONNX | w8a16 | Snapdragon® X Elite | 489.095 ms | 129 - 129 MB | NPU
73
- | Video-MAE | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 447.917 ms | 5 - 1858 MB | NPU
74
- | Video-MAE | ONNX | w8a16 | Qualcomm® QCS6490 | 9057.031 ms | 363 - 381 MB | CPU
75
- | Video-MAE | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 473.026 ms | 0 - 160 MB | NPU
76
- | Video-MAE | ONNX | w8a16 | Qualcomm® QCM6690 | 4452.468 ms | 326 - 338 MB | CPU
77
- | Video-MAE | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 278.713 ms | 0 - 1538 MB | NPU
78
- | Video-MAE | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 4377.307 ms | 376 - 388 MB | CPU
79
- | Video-MAE | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 323.42 ms | 5 - 1549 MB | NPU
80
- | Video-MAE | QNN_DLC | float | Snapdragon® X2 Elite | 249.277 ms | 9 - 9 MB | NPU
81
- | Video-MAE | QNN_DLC | float | Snapdragon® X Elite | 1308.348 ms | 9 - 9 MB | NPU
82
- | Video-MAE | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 414.95 ms | 9 - 1110 MB | NPU
83
- | Video-MAE | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 2499.681 ms | 1 - 889 MB | NPU
84
- | Video-MAE | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 1249.753 ms | 9 - 12 MB | NPU
85
- | Video-MAE | QNN_DLC | float | Qualcomm® SA8775P | 6237.053 ms | 1 - 889 MB | NPU
86
- | Video-MAE | QNN_DLC | float | Qualcomm® QCS9075 | 519.754 ms | 9 - 20 MB | NPU
87
- | Video-MAE | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 585.147 ms | 9 - 1025 MB | NPU
88
- | Video-MAE | QNN_DLC | float | Qualcomm® SA7255P | 2499.681 ms | 1 - 889 MB | NPU
89
- | Video-MAE | QNN_DLC | float | Qualcomm® SA8295P | 655.834 ms | 0 - 843 MB | NPU
90
- | Video-MAE | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 307.435 ms | 9 - 934 MB | NPU
91
- | Video-MAE | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 322.365 ms | 9 - 943 MB | NPU
92
- | Video-MAE | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 103.807 ms | 0 - 1114 MB | NPU
93
- | Video-MAE | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 428.414 ms | 0 - 918 MB | NPU
94
- | Video-MAE | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 141.755 ms | 0 - 3 MB | NPU
95
- | Video-MAE | TFLITE | float | Qualcomm® SA8775P | 164.5 ms | 0 - 921 MB | NPU
96
- | Video-MAE | TFLITE | float | Qualcomm® QCS9075 | 176.822 ms | 0 - 207 MB | NPU
97
- | Video-MAE | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 302.041 ms | 0 - 1056 MB | NPU
98
- | Video-MAE | TFLITE | float | Qualcomm® SA7255P | 428.414 ms | 0 - 918 MB | NPU
99
- | Video-MAE | TFLITE | float | Qualcomm® SA8295P | 216.681 ms | 0 - 872 MB | NPU
100
- | Video-MAE | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 81.01 ms | 0 - 923 MB | NPU
101
- | Video-MAE | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 63.809 ms | 0 - 941 MB | NPU
102
 
103
  ## License
104
  * The license for the original implementation of Video-MAE can be found
 
15
  Video MAE (Masked Auto Encoder) is a network for doing video classification that uses the ViT (Vision Transformer) backbone.
16
 
17
  This is based on the implementation of Video-MAE found [here](https://github.com/MCG-NJU/VideoMAE).
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/tree/v0.49.1/qai_hub_models/models/video_mae) 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
 
 
28
 
29
  | Runtime | Precision | Chipset | SDK Versions | Download |
30
  |---|---|---|---|---|
31
+ | ONNX | float | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.49.1/video_mae-onnx-float.zip)
32
+ | ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.49.1/video_mae-onnx-w8a16.zip)
33
+ | QNN_DLC | float | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.49.1/video_mae-qnn_dlc-float.zip)
34
+ | QNN_DLC | w8a16 | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.49.1/video_mae-qnn_dlc-w8a16.zip)
35
+ | TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/releases/v0.49.1/video_mae-tflite-float.zip)
36
 
37
  For more device-specific assets and performance metrics, visit **[Video-MAE on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/video_mae)**.
38
 
39
 
40
  ### Option 2: Export with Custom Configurations
41
 
42
+ Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/tree/v0.49.1/qai_hub_models/models/video_mae) Python library to compile and export the model with your own:
43
  - Custom weights (e.g., fine-tuned checkpoints)
44
  - Custom input shapes
45
  - Target device and runtime configurations
46
 
47
  This option is ideal if you need to customize the model beyond the default configuration provided here.
48
 
49
+ See our repository for [Video-MAE on GitHub](https://github.com/qualcomm/ai-hub-models/tree/v0.49.1/qai_hub_models/models/video_mae) for usage instructions.
50
 
51
  ## Model Details
52
 
 
61
  ## Performance Summary
62
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
63
  |---|---|---|---|---|---|---
64
+ | Video-MAE | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 445.683 ms | 9 - 1065 MB | NPU
65
+ | Video-MAE | ONNX | float | Snapdragon® X2 Elite | 464.546 ms | 188 - 188 MB | NPU
66
+ | Video-MAE | ONNX | float | Snapdragon® X Elite | 668.505 ms | 188 - 188 MB | NPU
67
+ | Video-MAE | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 387.58 ms | 2 - 1237 MB | NPU
68
+ | Video-MAE | ONNX | float | Qualcomm® QCS8550 (Proxy) | 645.578 ms | 0 - 219 MB | NPU
69
+ | Video-MAE | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 394.065 ms | 3 - 981 MB | NPU
70
+ | Video-MAE | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 324.004 ms | 5 - 1549 MB | NPU
71
+ | Video-MAE | ONNX | w8a16 | Snapdragon® X2 Elite | 323.982 ms | 123 - 123 MB | NPU
72
+ | Video-MAE | ONNX | w8a16 | Snapdragon® X Elite | 489.248 ms | 129 - 129 MB | NPU
73
+ | Video-MAE | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 447.019 ms | 5 - 1862 MB | NPU
74
+ | Video-MAE | ONNX | w8a16 | Qualcomm® QCS6490 | 9030.378 ms | 368 - 385 MB | CPU
75
+ | Video-MAE | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 474.16 ms | 0 - 166 MB | NPU
76
+ | Video-MAE | ONNX | w8a16 | Qualcomm® QCM6690 | 4430.795 ms | 331 - 343 MB | CPU
77
+ | Video-MAE | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 277.84 ms | 1 - 1538 MB | NPU
78
+ | Video-MAE | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 4366.928 ms | 332 - 345 MB | CPU
79
+ | Video-MAE | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 322.542 ms | 9 - 943 MB | NPU
80
+ | Video-MAE | QNN_DLC | float | Snapdragon® X2 Elite | 250.291 ms | 9 - 9 MB | NPU
81
+ | Video-MAE | QNN_DLC | float | Snapdragon® X Elite | 1308.489 ms | 9 - 9 MB | NPU
82
+ | Video-MAE | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 414.552 ms | 8 - 1110 MB | NPU
83
+ | Video-MAE | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 2501.763 ms | 2 - 891 MB | NPU
84
+ | Video-MAE | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 1250.286 ms | 9 - 699 MB | NPU
85
+ | Video-MAE | QNN_DLC | float | Qualcomm® SA8775P | 1300.648 ms | 0 - 888 MB | NPU
86
+ | Video-MAE | QNN_DLC | float | Qualcomm® QCS9075 | 520.331 ms | 11 - 22 MB | NPU
87
+ | Video-MAE | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 583.27 ms | 9 - 1024 MB | NPU
88
+ | Video-MAE | QNN_DLC | float | Qualcomm® SA7255P | 2501.763 ms | 2 - 891 MB | NPU
89
+ | Video-MAE | QNN_DLC | float | Qualcomm® SA8295P | 655.984 ms | 0 - 842 MB | NPU
90
+ | Video-MAE | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 307.095 ms | 0 - 925 MB | NPU
91
+ | Video-MAE | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 63.87 ms | 0 - 940 MB | NPU
92
+ | Video-MAE | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 104.056 ms | 0 - 1106 MB | NPU
93
+ | Video-MAE | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 428.321 ms | 0 - 917 MB | NPU
94
+ | Video-MAE | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 142.102 ms | 0 - 3 MB | NPU
95
+ | Video-MAE | TFLITE | float | Qualcomm® SA8775P | 164.352 ms | 0 - 921 MB | NPU
96
+ | Video-MAE | TFLITE | float | Qualcomm® QCS9075 | 176.877 ms | 0 - 207 MB | NPU
97
+ | Video-MAE | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 302.538 ms | 0 - 1060 MB | NPU
98
+ | Video-MAE | TFLITE | float | Qualcomm® SA7255P | 428.321 ms | 0 - 917 MB | NPU
99
+ | Video-MAE | TFLITE | float | Qualcomm® SA8295P | 216.712 ms | 0 - 872 MB | NPU
100
+ | Video-MAE | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 80.576 ms | 0 - 922 MB | NPU
 
101
 
102
  ## License
103
  * The license for the original implementation of Video-MAE can be found