Hugging Face's logo

qualcomm
/
CREStereo

Depth Estimation
PyTorch
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
dc3a473
·
verified ·
1 Parent(s): 4c688df

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 +46 -46
  2. release_assets.json +4 -4
README.md CHANGED
@@ -14,7 +14,7 @@ pipeline_tag: depth-estimation
14
  CREStereo (Cascaded Recurrent Network with Adaptive Correlation) is a CVPR 2022 Oral paper that achieves state-of-the-art stereo matching accuracy.
15
 
16
  This is based on the implementation of CREStereo found [here](https://github.com/ibaiGorordo/CREStereo-Pytorch).
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/crestereo) 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 | ONNX Runtime 1.25.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.57.3/crestereo-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/crestereo/releases/v0.57.3/crestereo-qnn_dlc-float.zip)
32
- | TFLITE | float | Universal | | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.57.3/crestereo-tflite-float.zip)
33
 
34
  For more device-specific assets and performance metrics, visit **[CREStereo on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/crestereo)**.
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/crestereo) 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 [CREStereo on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.57.3/src/qai_hub_models/models/crestereo) for usage instructions.
47
 
48
  ## Model Details
49
 
@@ -60,48 +60,48 @@ See our repository for [CREStereo on GitHub](https://github.com/qualcomm/ai-hub-
60
  ## Performance Summary
61
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
62
  |---|---|---|---|---|---|---
63
- | CREStereo | ONNX | float | Snapdragon® X2 Elite | 1705.182 ms | 184 - 184 MB | CPU
64
- | CREStereo | ONNX | float | Snapdragon® X Elite | 13237.424 ms | 182 - 182 MB | CPU
65
- | CREStereo | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 4818.812 ms | 35 - 55 MB | CPU
66
- | CREStereo | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 5936.471 ms | 62 - 88 MB | CPU
67
- | CREStereo | ONNX | float | Qualcomm® QCS8550 (Proxy) | 6053.112 ms | 38 - 42 MB | CPU
68
- | CREStereo | ONNX | float | Qualcomm® QCS8450 | 5936.471 ms | 62 - 88 MB | CPU
69
- | CREStereo | ONNX | float | Snapdragon® 8 Elite Mobile | 3866.958 ms | 9 - 25 MB | CPU
70
- | CREStereo | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 3689.841 ms | 22 - 39 MB | CPU
71
- | CREStereo | ONNX | float | Qualcomm® QCS9075 | 3843.146 ms | 71 - 73 MB | CPU
72
- | CREStereo | ONNX | float | Qualcomm® QCS8750 | 3866.958 ms | 9 - 25 MB | CPU
73
- | CREStereo | ONNX | float | Qualcomm® QCS7181 | 13237.424 ms | 182 - 182 MB | CPU
74
- | CREStereo | QNN_DLC | float | Snapdragon® X2 Elite | 2763.679 ms | 106 - 106 MB | CPU
75
- | CREStereo | QNN_DLC | float | Snapdragon® X Elite | 8026.947 ms | 106 - 106 MB | CPU
76
- | CREStereo | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 4840.298 ms | 124 - 202 MB | CPU
77
- | CREStereo | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 7045.308 ms | 150 - 231 MB | CPU
78
- | CREStereo | QNN_DLC | float | Qualcomm® QCS8275 | 11552.702 ms | 136 - 208 MB | CPU
79
- | CREStereo | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 5647.429 ms | 138 - 174 MB | CPU
80
- | CREStereo | QNN_DLC | float | Qualcomm® SA8775P | 7101.249 ms | 136 - 209 MB | CPU
81
- | CREStereo | QNN_DLC | float | Qualcomm® SA8650P | 7101.249 ms | 136 - 209 MB | CPU
82
- | CREStereo | QNN_DLC | float | Qualcomm® SA8255P | 7101.249 ms | 136 - 209 MB | CPU
83
- | CREStereo | QNN_DLC | float | Qualcomm® QCS8450 | 7045.308 ms | 150 - 231 MB | CPU
84
- | CREStereo | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 3499.965 ms | 126 - 203 MB | CPU
85
- | CREStereo | QNN_DLC | float | Qualcomm® SA8295P | 5651.013 ms | 143 - 215 MB | CPU
86
- | CREStereo | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 2882.637 ms | 109 - 187 MB | CPU
87
- | CREStereo | QNN_DLC | float | Qualcomm® SA7255P | 11552.702 ms | 136 - 208 MB | CPU
88
- | CREStereo | QNN_DLC | float | Qualcomm® QCS9075 | 7306.935 ms | 308 - 1207 MB | CPU
89
- | CREStereo | QNN_DLC | float | Qualcomm® QCS8750 | 3499.965 ms | 126 - 203 MB | CPU
90
- | CREStereo | QNN_DLC | float | Qualcomm® QCS7181 | 8026.947 ms | 106 - 106 MB | CPU
91
- | CREStereo | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 4466.995 ms | 39 - 65 MB | CPU
92
  | CREStereo | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 6864.416 ms | 41 - 73 MB | CPU
93
- | CREStereo | TFLITE | float | Qualcomm® QCS8275 | 7076.53 ms | 41 - 65 MB | CPU
94
- | CREStereo | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 5054.19 ms | 37 - 49 MB | CPU
95
- | CREStereo | TFLITE | float | Qualcomm® SA8775P | 5630.969 ms | 41 - 65 MB | CPU
96
- | CREStereo | TFLITE | float | Qualcomm® SA8650P | 5630.969 ms | 41 - 65 MB | CPU
97
- | CREStereo | TFLITE | float | Qualcomm® SA8255P | 5630.969 ms | 41 - 65 MB | CPU
98
  | CREStereo | TFLITE | float | Qualcomm® QCS8450 | 6864.416 ms | 41 - 73 MB | CPU
99
- | CREStereo | TFLITE | float | Snapdragon® 8 Elite Mobile | 2875.056 ms | 39 - 66 MB | CPU
100
- | CREStereo | TFLITE | float | Qualcomm® SA8295P | 3718.184 ms | 40 - 64 MB | CPU
101
- | CREStereo | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 1097.651 ms | 30 - 60 MB | CPU
102
- | CREStereo | TFLITE | float | Qualcomm® SA7255P | 7076.53 ms | 41 - 65 MB | CPU
103
- | CREStereo | TFLITE | float | Qualcomm® QCS9075 | 5679.714 ms | 40 - 358 MB | CPU
104
- | CREStereo | TFLITE | float | Qualcomm® QCS8750 | 2875.056 ms | 39 - 66 MB | CPU
105
 
106
  ## License
107
  * The license for the original implementation of CREStereo can be found
 
14
  CREStereo (Cascaded Recurrent Network with Adaptive Correlation) is a CVPR 2022 Oral paper that achieves state-of-the-art stereo matching accuracy.
15
 
16
  This is based on the implementation of CREStereo found [here](https://github.com/ibaiGorordo/CREStereo-Pytorch).
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/crestereo) 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 | ONNX Runtime 1.25.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.58.0/crestereo-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/crestereo/releases/v0.58.0/crestereo-qnn_dlc-float.zip)
32
+ | TFLITE | float | Universal | | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.58.0/crestereo-tflite-float.zip)
33
 
34
  For more device-specific assets and performance metrics, visit **[CREStereo on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/crestereo)**.
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/crestereo) 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 [CREStereo on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.58.0/src/qai_hub_models/models/crestereo) for usage instructions.
47
 
48
  ## Model Details
49
 
 
60
  ## Performance Summary
61
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
62
  |---|---|---|---|---|---|---
63
+ | CREStereo | ONNX | float | Snapdragon® X2 Elite | 91.676 ms | 2 - 2 MB | NPU
64
+ | CREStereo | ONNX | float | Snapdragon® X Elite | 211.061 ms | 23 - 23 MB | NPU
65
+ | CREStereo | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 146.939 ms | 3 - 1520 MB | NPU
66
+ | CREStereo | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 357.129 ms | 4 - 1511 MB | NPU
67
+ | CREStereo | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 206.27 ms | 0 - 31 MB | NPU
68
+ | CREStereo | ONNX | float | Qualcomm® QCS8450 | 357.129 ms | 4 - 1511 MB | NPU
69
+ | CREStereo | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 312.554 ms | 2 - 5 MB | NPU
70
+ | CREStereo | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 88.465 ms | 2 - 937 MB | NPU
71
+ | CREStereo | ONNX | float | Snapdragon® 8 Elite Mobile | 116.301 ms | 3 - 1155 MB | NPU
72
+ | CREStereo | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 116.301 ms | 3 - 1155 MB | NPU
73
+ | CREStereo | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 211.061 ms | 23 - 23 MB | NPU
74
+ | CREStereo | QNN_DLC | float | Snapdragon® X2 Elite | 90.738 ms | 2 - 2 MB | NPU
75
+ | CREStereo | QNN_DLC | float | Snapdragon® X Elite | 209.682 ms | 2 - 2 MB | NPU
76
+ | CREStereo | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 149.943 ms | 3 - 1528 MB | NPU
77
+ | CREStereo | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 358.307 ms | 2 - 1841 MB | NPU
78
+ | CREStereo | QNN_DLC | float | Qualcomm® QCS8275 | 807.424 ms | 3 - 1186 MB | NPU
79
+ | CREStereo | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 205.514 ms | 3 - 5 MB | NPU
80
+ | CREStereo | QNN_DLC | float | Qualcomm® SA8775P | 273.08 ms | 3 - 1403 MB | NPU
81
+ | CREStereo | QNN_DLC | float | Qualcomm® SA8650P | 273.08 ms | 3 - 1403 MB | NPU
82
+ | CREStereo | QNN_DLC | float | Qualcomm® SA8255P | 273.08 ms | 3 - 1403 MB | NPU
83
+ | CREStereo | QNN_DLC | float | Qualcomm® QCS8450 | 358.307 ms | 2 - 1841 MB | NPU
84
+ | CREStereo | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 272.205 ms | 2 - 6 MB | NPU
85
+ | CREStereo | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 81.623 ms | 1 - 786 MB | NPU
86
+ | CREStereo | QNN_DLC | float | Qualcomm® SA7255P | 807.424 ms | 3 - 1186 MB | NPU
87
+ | CREStereo | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 116.402 ms | 2 - 1240 MB | NPU
88
+ | CREStereo | QNN_DLC | float | Qualcomm® SA8295P | 317.035 ms | 2 - 1390 MB | NPU
89
+ | CREStereo | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 116.402 ms | 2 - 1240 MB | NPU
90
+ | CREStereo | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 209.682 ms | 2 - 2 MB | NPU
91
+ | CREStereo | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 153.2 ms | 0 - 2171 MB | NPU
92
  | CREStereo | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 6864.416 ms | 41 - 73 MB | CPU
93
+ | CREStereo | TFLITE | float | Qualcomm® QCS8275 | 813.553 ms | 1 - 1599 MB | NPU
94
+ | CREStereo | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 214.732 ms | 0 - 7 MB | NPU
95
+ | CREStereo | TFLITE | float | Qualcomm® SA8775P | 282.035 ms | 1 - 1596 MB | NPU
96
+ | CREStereo | TFLITE | float | Qualcomm® SA8650P | 282.035 ms | 1 - 1596 MB | NPU
97
+ | CREStereo | TFLITE | float | Qualcomm® SA8255P | 282.035 ms | 1 - 1596 MB | NPU
98
  | CREStereo | TFLITE | float | Qualcomm® QCS8450 | 6864.416 ms | 41 - 73 MB | CPU
99
+ | CREStereo | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 281.095 ms | 0 - 44 MB | NPU
100
+ | CREStereo | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 89.079 ms | 0 - 1270 MB | NPU
101
+ | CREStereo | TFLITE | float | Qualcomm® SA7255P | 813.553 ms | 1 - 1599 MB | NPU
102
+ | CREStereo | TFLITE | float | Snapdragon® 8 Elite Mobile | 122.971 ms | 1 - 1652 MB | NPU
103
+ | CREStereo | TFLITE | float | Qualcomm® SA8295P | 357.765 ms | 1 - 1858 MB | NPU
104
+ | CREStereo | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 122.971 ms | 1 - 1652 MB | NPU
105
 
106
  ## License
107
  * The license for the original implementation of CREStereo can be found
release_assets.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "version": "0.57.3",
3
  "precisions": {
4
  "float": {
5
  "universal_assets": {
@@ -7,19 +7,19 @@
7
  "tool_versions": {
8
  "litert": "1.4.4"
9
  },
10
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.57.3/crestereo-tflite-float.zip"
11
  },
12
  "qnn_dlc": {
13
  "tool_versions": {
14
  "qairt": "2.45.0.260326154327"
15
  },
16
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.57.3/crestereo-qnn_dlc-float.zip"
17
  },
18
  "onnx": {
19
  "tool_versions": {
20
  "onnx_runtime": "1.25.0"
21
  },
22
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.57.3/crestereo-onnx-float.zip"
23
  }
24
  }
25
  }
 
1
  {
2
+ "version": "0.58.0",
3
  "precisions": {
4
  "float": {
5
  "universal_assets": {
 
7
  "tool_versions": {
8
  "litert": "1.4.4"
9
  },
10
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.58.0/crestereo-tflite-float.zip"
11
  },
12
  "qnn_dlc": {
13
  "tool_versions": {
14
  "qairt": "2.45.0.260326154327"
15
  },
16
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.58.0/crestereo-qnn_dlc-float.zip"
17
  },
18
  "onnx": {
19
  "tool_versions": {
20
  "onnx_runtime": "1.25.0"
21
  },
22
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/crestereo/releases/v0.58.0/crestereo-onnx-float.zip"
23
  }
24
  }
25
  }