Hugging Face's logo

qualcomm
/
PointNet

Image Segmentation
PyTorch
bu_auto
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
af51006
·
verified ·
1 Parent(s): b9f6e20

v0.60.0

Browse files

See https://github.com/qualcomm/ai-hub-models/releases/v0.60.0 for changelog.

Files changed (2) hide show
  1. README.md +48 -45
  2. release_assets.json +4 -4
README.md CHANGED
@@ -14,7 +14,7 @@ pipeline_tag: image-segmentation
14
 
15
  PointNet is a pioneering neural network architecture designed to directly consume unordered point cloud data for tasks such as classification and segmentation. It learns spatial features from raw 3D points without requiring voxelization or image projections.
16
 
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/v0.59.0/src/qai_hub_models/models/pointnet) 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
 
@@ -27,78 +27,81 @@ 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 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pointnet/releases/v0.59.0/pointnet-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/pointnet/releases/v0.59.0/pointnet-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/pointnet/releases/v0.59.0/pointnet-tflite-float.zip)
33
 
34
  For more device-specific assets and performance metrics, visit **[PointNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/pointnet)**.
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/v0.59.0/src/qai_hub_models/models/pointnet) 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 [PointNet on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.59.0/src/qai_hub_models/models/pointnet) for usage instructions.
47
 
48
  ## Model Details
49
 
50
  **Model Type:** Model_use_case.semantic_segmentation
51
 
52
  **Model Stats:**
53
- - Model checkpoint: save
54
  - Input resolution: 1x3x1024
 
55
  - Model size: 13.2 MB
56
 
57
  ## Performance Summary
58
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
59
  |---|---|---|---|---|---|---
60
- | PointNet | ONNX | float | Snapdragon® X2 Elite | 0.302 ms | 1 - 1 MB | NPU
61
- | PointNet | ONNX | float | Snapdragon® X Elite | 0.66 ms | 7 - 7 MB | NPU
62
- | PointNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.401 ms | 0 - 43 MB | NPU
63
- | PointNet | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 0.865 ms | 0 - 49 MB | NPU
64
- | PointNet | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.646 ms | 0 - 85 MB | NPU
65
- | PointNet | ONNX | float | Qualcomm® QCS8450 | 0.865 ms | 0 - 49 MB | NPU
66
- | PointNet | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 0.804 ms | 0 - 3 MB | NPU
67
- | PointNet | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.66 ms | 7 - 7 MB | NPU
 
68
  | PointNet | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.42 ms | 0 - 28 MB | NPU
69
  | PointNet | ONNX | float | Snapdragon® 8 Elite Mobile | 0.42 ms | 0 - 28 MB | NPU
70
- | PointNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.324 ms | 0 - 23 MB | NPU
71
- | PointNet | QNN_DLC | float | Snapdragon® X2 Elite | 0.412 ms | 0 - 0 MB | NPU
72
  | PointNet | QNN_DLC | float | Snapdragon® X Elite | 0.771 ms | 1 - 1 MB | NPU
73
- | PointNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 0.418 ms | 0 - 42 MB | NPU
74
- | PointNet | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 0.865 ms | 0 - 43 MB | NPU
75
- | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8275 | 1.861 ms | 0 - 24 MB | NPU
76
- | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.661 ms | 0 - 1 MB | NPU
77
- | PointNet | QNN_DLC | float | Qualcomm® SA8775P | 0.917 ms | 0 - 25 MB | NPU
78
- | PointNet | QNN_DLC | float | Qualcomm® SA8650P | 0.917 ms | 0 - 25 MB | NPU
79
- | PointNet | QNN_DLC | float | Qualcomm® SA8255P | 0.917 ms | 0 - 25 MB | NPU
80
- | PointNet | QNN_DLC | float | Qualcomm® QCS8450 | 0.865 ms | 0 - 43 MB | NPU
81
- | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 0.812 ms | 0 - 2 MB | NPU
 
82
  | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.771 ms | 1 - 1 MB | NPU
83
- | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 0.482 ms | 0 - 23 MB | NPU
84
- | PointNet | QNN_DLC | float | Qualcomm® SA7255P | 1.861 ms | 0 - 24 MB | NPU
85
- | PointNet | QNN_DLC | float | Qualcomm® SA8295P | 1.108 ms | 0 - 21 MB | NPU
86
- | PointNet | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 0.482 ms | 0 - 23 MB | NPU
87
- | PointNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.324 ms | 0 - 23 MB | NPU
88
- | PointNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 0.409 ms | 0 - 41 MB | NPU
89
- | PointNet | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 0.86 ms | 0 - 47 MB | NPU
90
- | PointNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8275 | 1.887 ms | 0 - 24 MB | NPU
91
- | PointNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.665 ms | 0 - 16 MB | NPU
92
- | PointNet | TFLITE | float | Qualcomm® SA8775P | 0.922 ms | 0 - 25 MB | NPU
93
- | PointNet | TFLITE | float | Qualcomm® SA8650P | 0.922 ms | 0 - 25 MB | NPU
94
- | PointNet | TFLITE | float | Qualcomm® SA8255P | 0.922 ms | 0 - 25 MB | NPU
95
- | PointNet | TFLITE | float | Qualcomm® QCS8450 | 0.86 ms | 0 - 47 MB | NPU
96
- | PointNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 0.816 ms | 0 - 8 MB | NPU
97
- | PointNet | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 0.483 ms | 0 - 24 MB | NPU
98
- | PointNet | TFLITE | float | Qualcomm® SA7255P | 1.887 ms | 0 - 24 MB | NPU
99
- | PointNet | TFLITE | float | Qualcomm® SA8295P | 1.12 ms | 0 - 21 MB | NPU
100
- | PointNet | TFLITE | float | Snapdragon® 8 Elite Mobile | 0.483 ms | 0 - 24 MB | NPU
101
- | PointNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.324 ms | 0 - 24 MB | NPU
 
102
 
103
  ## License
104
  * The license for the original implementation of PointNet can be found
 
14
 
15
  PointNet is a pioneering neural network architecture designed to directly consume unordered point cloud data for tasks such as classification and segmentation. It learns spatial features from raw 3D points without requiring voxelization or image projections.
16
 
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/v0.60.0/src/qai_hub_models/models/pointnet) 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 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pointnet/releases/v0.60.0/pointnet-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/pointnet/releases/v0.60.0/pointnet-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/pointnet/releases/v0.60.0/pointnet-tflite-float.zip)
33
 
34
  For more device-specific assets and performance metrics, visit **[PointNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/pointnet)**.
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/v0.60.0/src/qai_hub_models/models/pointnet) 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 [PointNet on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/pointnet) for usage instructions.
47
 
48
  ## Model Details
49
 
50
  **Model Type:** Model_use_case.semantic_segmentation
51
 
52
  **Model Stats:**
 
53
  - Input resolution: 1x3x1024
54
+ - Model checkpoint: save
55
  - Model size: 13.2 MB
56
 
57
  ## Performance Summary
58
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
59
  |---|---|---|---|---|---|---
60
+ | PointNet | ONNX | float | Snapdragon® X2 Elite | 0.303 ms | 1 - 1 MB | NPU
61
+ | PointNet | ONNX | float | Snapdragon® X Elite | 0.661 ms | 7 - 7 MB | NPU
62
+ | PointNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.403 ms | 0 - 42 MB | NPU
63
+ | PointNet | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 0.865 ms | 0 - 47 MB | NPU
64
+ | PointNet | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 0.737 ms | 0 - 3 MB | NPU
65
+ | PointNet | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.644 ms | 0 - 9 MB | NPU
66
+ | PointNet | ONNX | float | Qualcomm® QCS8450 | 0.865 ms | 0 - 47 MB | NPU
67
+ | PointNet | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 0.806 ms | 0 - 3 MB | NPU
68
+ | PointNet | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.661 ms | 7 - 7 MB | NPU
69
  | PointNet | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.42 ms | 0 - 28 MB | NPU
70
  | PointNet | ONNX | float | Snapdragon® 8 Elite Mobile | 0.42 ms | 0 - 28 MB | NPU
71
+ | PointNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.325 ms | 0 - 23 MB | NPU
72
+ | PointNet | QNN_DLC | float | Snapdragon® X2 Elite | 0.408 ms | 0 - 0 MB | NPU
73
  | PointNet | QNN_DLC | float | Snapdragon® X Elite | 0.771 ms | 1 - 1 MB | NPU
74
+ | PointNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 0.41 ms | 0 - 42 MB | NPU
75
+ | PointNet | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 0.864 ms | 0 - 47 MB | NPU
76
+ | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 0.738 ms | 0 - 2 MB | NPU
77
+ | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 1.866 ms | 0 - 24 MB | NPU
78
+ | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.659 ms | 0 - 3 MB | NPU
79
+ | PointNet | QNN_DLC | float | Qualcomm® SA8775P | 0.916 ms | 0 - 25 MB | NPU
80
+ | PointNet | QNN_DLC | float | Qualcomm® SA8650P | 0.916 ms | 0 - 25 MB | NPU
81
+ | PointNet | QNN_DLC | float | Qualcomm® SA8255P | 0.916 ms | 0 - 25 MB | NPU
82
+ | PointNet | QNN_DLC | float | Qualcomm® QCS8450 | 0.864 ms | 0 - 47 MB | NPU
83
+ | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 0.815 ms | 2 - 4 MB | NPU
84
  | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.771 ms | 1 - 1 MB | NPU
85
+ | PointNet | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 0.481 ms | 0 - 23 MB | NPU
86
+ | PointNet | QNN_DLC | float | Qualcomm® SA7255P | 1.866 ms | 0 - 24 MB | NPU
87
+ | PointNet | QNN_DLC | float | Qualcomm® SA8295P | 1.107 ms | 0 - 21 MB | NPU
88
+ | PointNet | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 0.481 ms | 0 - 23 MB | NPU
89
+ | PointNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.325 ms | 0 - 24 MB | NPU
90
+ | PointNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 0.41 ms | 0 - 41 MB | NPU
91
+ | PointNet | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 0.861 ms | 0 - 47 MB | NPU
92
+ | PointNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 0.743 ms | 0 - 9 MB | NPU
93
+ | PointNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 1.871 ms | 0 - 25 MB | NPU
94
+ | PointNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.65 ms | 0 - 1 MB | NPU
95
+ | PointNet | TFLITE | float | Qualcomm® SA8775P | 0.921 ms | 0 - 25 MB | NPU
96
+ | PointNet | TFLITE | float | Qualcomm® SA8650P | 0.921 ms | 0 - 25 MB | NPU
97
+ | PointNet | TFLITE | float | Qualcomm® SA8255P | 0.921 ms | 0 - 25 MB | NPU
98
+ | PointNet | TFLITE | float | Qualcomm® QCS8450 | 0.861 ms | 0 - 47 MB | NPU
99
+ | PointNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 0.825 ms | 0 - 8 MB | NPU
100
+ | PointNet | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 0.483 ms | 0 - 28 MB | NPU
101
+ | PointNet | TFLITE | float | Qualcomm® SA7255P | 1.871 ms | 0 - 25 MB | NPU
102
+ | PointNet | TFLITE | float | Qualcomm® SA8295P | 1.138 ms | 0 - 22 MB | NPU
103
+ | PointNet | TFLITE | float | Snapdragon® 8 Elite Mobile | 0.483 ms | 0 - 28 MB | NPU
104
+ | PointNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.32 ms | 0 - 25 MB | NPU
105
 
106
  ## License
107
  * The license for the original implementation of PointNet can be found
release_assets.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "version": "0.59.0",
3
  "precisions": {
4
  "float": {
5
  "universal_assets": {
@@ -8,19 +8,19 @@
8
  "qairt": "2.45.0.260326154327",
9
  "onnx_runtime": "1.27.1"
10
  },
11
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pointnet/releases/v0.59.0/pointnet-onnx-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/pointnet/releases/v0.59.0/pointnet-qnn_dlc-float.zip"
18
  },
19
  "tflite": {
20
  "tool_versions": {
21
  "qairt": "2.45.0.260326154327"
22
  },
23
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pointnet/releases/v0.59.0/pointnet-tflite-float.zip"
24
  }
25
  }
26
  }
 
1
  {
2
+ "version": "0.60.0",
3
  "precisions": {
4
  "float": {
5
  "universal_assets": {
 
8
  "qairt": "2.45.0.260326154327",
9
  "onnx_runtime": "1.27.1"
10
  },
11
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pointnet/releases/v0.60.0/pointnet-onnx-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/pointnet/releases/v0.60.0/pointnet-qnn_dlc-float.zip"
18
  },
19
  "tflite": {
20
  "tool_versions": {
21
  "qairt": "2.45.0.260326154327"
22
  },
23
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/pointnet/releases/v0.60.0/pointnet-tflite-float.zip"
24
  }
25
  }
26
  }