Hugging Face's logo

qualcomm
/
PSPNet

Image Segmentation
PyTorch
bu_auto
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
d5f9d91
·
verified ·
1 Parent(s): 6fbffe7

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 +53 -50
  2. release_assets.json +4 -4
README.md CHANGED
@@ -14,7 +14,7 @@ pipeline_tag: image-segmentation
14
 
15
  PSPNet (Pyramid Scene Parsing Network) is a semantic segmentation model that captures global context information by applying pyramid pooling modules. It is designed to improve scene understanding by aggregating contextual features at multiple scales.
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/pspnet) 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
 
@@ -27,79 +27,82 @@ 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/pspnet/releases/v0.59.0/pspnet-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/pspnet/releases/v0.59.0/pspnet-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/pspnet/releases/v0.59.0/pspnet-tflite-float.zip)
33
 
34
  For more device-specific assets and performance metrics, visit **[PSPNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/pspnet)**.
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/pspnet) 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 [PSPNet on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.59.0/src/qai_hub_models/models/pspnet) for usage instructions.
47
 
48
  ## Model Details
49
 
50
  **Model Type:** Model_use_case.semantic_segmentation
51
 
52
  **Model Stats:**
53
- - Model checkpoint: pspnet101_ade20k.pth
54
  - Input resolution: 1x3x473x473
55
- - Number of parameters: 65.7M
56
  - Model size (float): 251 MB
 
57
 
58
  ## Performance Summary
59
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
60
  |---|---|---|---|---|---|---
61
- | PSPNet | ONNX | float | Snapdragon® X2 Elite | 835.974 ms | 528 - 528 MB | NPU
62
- | PSPNet | ONNX | float | Snapdragon® X Elite | 1334.498 ms | 267 - 267 MB | NPU
63
- | PSPNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 956.615 ms | 142 - 1994 MB | NPU
64
- | PSPNet | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 3276.923 ms | 34 - 887 MB | NPU
65
- | PSPNet | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 1164.165 ms | 0 - 160 MB | NPU
66
- | PSPNet | ONNX | float | Qualcomm® QCS8450 | 3276.923 ms | 34 - 887 MB | NPU
67
- | PSPNet | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 1417.954 ms | 124 - 130 MB | NPU
68
- | PSPNet | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 1334.498 ms | 267 - 267 MB | NPU
69
- | PSPNet | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 649.02 ms | 133 - 1599 MB | NPU
70
- | PSPNet | ONNX | float | Snapdragon® 8 Elite Mobile | 649.02 ms | 133 - 1599 MB | NPU
71
- | PSPNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 718.907 ms | 244 - 1842 MB | NPU
72
- | PSPNet | QNN_DLC | float | Snapdragon® X2 Elite | 2497.872 ms | 3 - 3 MB | NPU
73
- | PSPNet | QNN_DLC | float | Snapdragon® X Elite | 2536.944 ms | 3 - 3 MB | NPU
74
- | PSPNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 1836.59 ms | 25 - 1690 MB | NPU
75
- | PSPNet | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 1595.526 ms | 0 - 851 MB | NPU
76
- | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8275 | 5274.515 ms | 13 - 1329 MB | NPU
77
- | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 2479.634 ms | 3 - 6 MB | NPU
78
- | PSPNet | QNN_DLC | float | Qualcomm® SA8775P | 2606.587 ms | 2 - 1319 MB | NPU
79
- | PSPNet | QNN_DLC | float | Qualcomm® SA8650P | 2606.587 ms | 2 - 1319 MB | NPU
80
- | PSPNet | QNN_DLC | float | Qualcomm® SA8255P | 2606.587 ms | 2 - 1319 MB | NPU
81
- | PSPNet | QNN_DLC | float | Qualcomm® QCS8450 | 1595.526 ms | 0 - 851 MB | NPU
82
- | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 2593.163 ms | 3 - 135 MB | NPU
83
- | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 2536.944 ms | 3 - 3 MB | NPU
84
- | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 2177.701 ms | 0 - 1311 MB | NPU
85
- | PSPNet | QNN_DLC | float | Qualcomm® SA7255P | 5274.515 ms | 13 - 1329 MB | NPU
86
- | PSPNet | QNN_DLC | float | Qualcomm® SA8295P | 1357.596 ms | 3 - 647 MB | NPU
87
- | PSPNet | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 2177.701 ms | 0 - 1311 MB | NPU
88
- | PSPNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 2340.9 ms | 0 - 1369 MB | NPU
89
- | PSPNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2104.97 ms | 1 - 1707 MB | NPU
90
- | PSPNet | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 1856.683 ms | 1 - 934 MB | NPU
91
- | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8275 | 5962.221 ms | 110 - 1507 MB | NPU
92
- | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 2829.444 ms | 0 - 4 MB | NPU
93
- | PSPNet | TFLITE | float | Qualcomm® SA8775P | 2955.535 ms | 111 - 1509 MB | NPU
94
- | PSPNet | TFLITE | float | Qualcomm® SA8650P | 2955.535 ms | 111 - 1509 MB | NPU
95
- | PSPNet | TFLITE | float | Qualcomm® SA8255P | 2955.535 ms | 111 - 1509 MB | NPU
96
- | PSPNet | TFLITE | float | Qualcomm® QCS8450 | 1856.683 ms | 1 - 934 MB | NPU
97
- | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 2938.747 ms | 64 - 339 MB | NPU
98
- | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 2172.868 ms | 1 - 1409 MB | NPU
99
- | PSPNet | TFLITE | float | Qualcomm® SA7255P | 5962.221 ms | 110 - 1507 MB | NPU
100
- | PSPNet | TFLITE | float | Qualcomm® SA8295P | 1414.274 ms | 128 - 837 MB | NPU
101
- | PSPNet | TFLITE | float | Snapdragon® 8 Elite Mobile | 2172.868 ms | 1 - 1409 MB | NPU
102
- | PSPNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2329.061 ms | 0 - 1451 MB | NPU
 
 
 
103
 
104
  ## License
105
  * The license for the original implementation of PSPNet can be found
 
14
 
15
  PSPNet (Pyramid Scene Parsing Network) is a semantic segmentation model that captures global context information by applying pyramid pooling modules. It is designed to improve scene understanding by aggregating contextual features at multiple scales.
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/pspnet) 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/pspnet/releases/v0.60.0/pspnet-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/pspnet/releases/v0.60.0/pspnet-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/pspnet/releases/v0.60.0/pspnet-tflite-float.zip)
33
 
34
  For more device-specific assets and performance metrics, visit **[PSPNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/pspnet)**.
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/pspnet) 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 [PSPNet on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/pspnet) for usage instructions.
47
 
48
  ## Model Details
49
 
50
  **Model Type:** Model_use_case.semantic_segmentation
51
 
52
  **Model Stats:**
 
53
  - Input resolution: 1x3x473x473
54
+ - Model checkpoint: pspnet101_ade20k.pth
55
  - Model size (float): 251 MB
56
+ - Number of parameters: 65.7M
57
 
58
  ## Performance Summary
59
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
60
  |---|---|---|---|---|---|---
61
+ | PSPNet | ONNX | float | Snapdragon® X2 Elite | 833.922 ms | 532 - 532 MB | NPU
62
+ | PSPNet | ONNX | float | Snapdragon® X Elite | 1338.221 ms | 267 - 267 MB | NPU
63
+ | PSPNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 956.163 ms | 207 - 2054 MB | NPU
64
+ | PSPNet | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 2237.393 ms | 33 - 887 MB | NPU
65
+ | PSPNet | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 1389.513 ms | 117 - 123 MB | NPU
66
+ | PSPNet | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 1166.652 ms | 0 - 160 MB | NPU
67
+ | PSPNet | ONNX | float | Qualcomm® QCS8450 | 2237.393 ms | 33 - 887 MB | NPU
68
+ | PSPNet | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 1684.754 ms | 114 - 120 MB | NPU
69
+ | PSPNet | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 1338.221 ms | 267 - 267 MB | NPU
70
+ | PSPNet | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 648.933 ms | 170 - 1644 MB | NPU
71
+ | PSPNet | ONNX | float | Snapdragon® 8 Elite Mobile | 648.933 ms | 170 - 1644 MB | NPU
72
+ | PSPNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 717.737 ms | 117 - 1713 MB | NPU
73
+ | PSPNet | QNN_DLC | float | Snapdragon® X2 Elite | 2505.903 ms | 3 - 3 MB | NPU
74
+ | PSPNet | QNN_DLC | float | Snapdragon® X Elite | 2540.633 ms | 3 - 3 MB | NPU
75
+ | PSPNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 1847.35 ms | 0 - 1654 MB | NPU
76
+ | PSPNet | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 1609.246 ms | 0 - 852 MB | NPU
77
+ | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 2531.233 ms | 3 - 136 MB | NPU
78
+ | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 5289.358 ms | 0 - 1305 MB | NPU
79
+ | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 2489.852 ms | 3 - 5 MB | NPU
80
+ | PSPNet | QNN_DLC | float | Qualcomm® SA8775P | 2605.883 ms | 1 - 1307 MB | NPU
81
+ | PSPNet | QNN_DLC | float | Qualcomm® SA8650P | 2605.883 ms | 1 - 1307 MB | NPU
82
+ | PSPNet | QNN_DLC | float | Qualcomm® SA8255P | 2605.883 ms | 1 - 1307 MB | NPU
83
+ | PSPNet | QNN_DLC | float | Qualcomm® QCS8450 | 1609.246 ms | 0 - 852 MB | NPU
84
+ | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 2598.163 ms | 3 - 135 MB | NPU
85
+ | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 2540.633 ms | 3 - 3 MB | NPU
86
+ | PSPNet | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 2168.941 ms | 0 - 1310 MB | NPU
87
+ | PSPNet | QNN_DLC | float | Qualcomm® SA7255P | 5289.358 ms | 0 - 1305 MB | NPU
88
+ | PSPNet | QNN_DLC | float | Qualcomm® SA8295P | 1361.805 ms | 3 - 647 MB | NPU
89
+ | PSPNet | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 2168.941 ms | 0 - 1310 MB | NPU
90
+ | PSPNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 2341.485 ms | 0 - 1363 MB | NPU
91
+ | PSPNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2105.151 ms | 42 - 1747 MB | NPU
92
+ | PSPNet | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 1861.665 ms | 129 - 1060 MB | NPU
93
+ | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 2879.024 ms | 3 - 278 MB | NPU
94
+ | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 5956.898 ms | 130 - 1528 MB | NPU
95
+ | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 2841.025 ms | 0 - 4 MB | NPU
96
+ | PSPNet | TFLITE | float | Qualcomm® SA8775P | 2954.21 ms | 123 - 1521 MB | NPU
97
+ | PSPNet | TFLITE | float | Qualcomm® SA8650P | 2954.21 ms | 123 - 1521 MB | NPU
98
+ | PSPNet | TFLITE | float | Qualcomm® SA8255P | 2954.21 ms | 123 - 1521 MB | NPU
99
+ | PSPNet | TFLITE | float | Qualcomm® QCS8450 | 1861.665 ms | 129 - 1060 MB | NPU
100
+ | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 2922.32 ms | 16 - 291 MB | NPU
101
+ | PSPNet | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 2165.438 ms | 1 - 1409 MB | NPU
102
+ | PSPNet | TFLITE | float | Qualcomm® SA7255P | 5956.898 ms | 130 - 1528 MB | NPU
103
+ | PSPNet | TFLITE | float | Qualcomm® SA8295P | 1421.541 ms | 71 - 780 MB | NPU
104
+ | PSPNet | TFLITE | float | Snapdragon® 8 Elite Mobile | 2165.438 ms | 1 - 1409 MB | NPU
105
+ | PSPNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2329.54 ms | 5 - 1456 MB | NPU
106
 
107
  ## License
108
  * The license for the original implementation of PSPNet 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/pspnet/releases/v0.59.0/pspnet-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/pspnet/releases/v0.59.0/pspnet-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/pspnet/releases/v0.59.0/pspnet-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/pspnet/releases/v0.60.0/pspnet-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/pspnet/releases/v0.60.0/pspnet-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/pspnet/releases/v0.60.0/pspnet-tflite-float.zip"
24
  }
25
  }
26
  }