Hugging Face's logo

qualcomm
/
DeepLabXception

Image Segmentation
PyTorch
LiteRT
bu_auto
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
0d3a78b
·
verified ·
1 Parent(s): 65c07fb

v0.36.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.36.0 for changelog.

Files changed (17) hide show
  1. .gitattributes +3 -0
  2. DEPLOYMENT_MODEL_LICENSE.pdf +3 -0
  3. DeepLabXception.dlc +3 -0
  4. DeepLabXception.onnx.zip +3 -0
  5. DeepLabXception.tflite +3 -0
  6. DeepLabXception_w8a8.dlc +3 -0
  7. DeepLabXception_w8a8.tflite +3 -0
  8. LICENSE +2 -0
  9. README.md +285 -0
  10. precompiled/qualcomm-qcs6490-proxy/DeepLabXception_w8a8.bin +3 -0
  11. precompiled/qualcomm-qcs6490-proxy/sdk_versions.yml +3 -0
  12. precompiled/qualcomm-snapdragon-x-elite/DeepLabXception.bin +3 -0
  13. precompiled/qualcomm-snapdragon-x-elite/DeepLabXception.onnx.zip +3 -0
  14. precompiled/qualcomm-snapdragon-x-elite/DeepLabXception_w8a8.bin +3 -0
  15. precompiled/qualcomm-snapdragon-x-elite/DeepLabXception_w8a8.onnx.zip +3 -0
  16. precompiled/qualcomm-snapdragon-x-elite/sdk_versions.yml +5 -0
  17. sdk_versions.yml +9 -0
.gitattributes CHANGED
@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ DEPLOYMENT_MODEL_LICENSE.pdf filter=lfs diff=lfs merge=lfs -text
37
+ DeepLabXception.dlc filter=lfs diff=lfs merge=lfs -text
38
+ DeepLabXception_w8a8.dlc filter=lfs diff=lfs merge=lfs -text
DEPLOYMENT_MODEL_LICENSE.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4409f93b0e82531303b3e10f52f1fdfb56467a25f05b7441c6bbd8bb8a64b42c
3
+ size 109629
DeepLabXception.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:775e6ab8807253a6cba1e99c5d9975952118c72e6baae50cd2de1e9647712473
3
+ size 164172444
DeepLabXception.onnx.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b5ca725f228799952cc74e5bc330c713871ae9290d91efea8fa5ddd85a93867
3
+ size 153177293
DeepLabXception.tflite ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6d1104cec0967e6cb11f4b969b77243406774bb1fbc66f9e7092363efe4e190
3
+ size 163888804
DeepLabXception_w8a8.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26257630cf1a1a718c992ab351434cc7681e531dc59dabba132ed75ef3c2deb9
3
+ size 47311188
DeepLabXception_w8a8.tflite ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b73d15b6dceaab252537ac65b4716e037f7512acd8717d4b02c03cc8d131da3e
3
+ size 43792240
LICENSE ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ The license of the original trained model can be found at https://github.com/pytorch/vision/blob/main/LICENSE.
2
+ The license for the deployable model files (.tflite, .onnx, .dlc, .bin, etc.) can be found in DEPLOYMENT_MODEL_LICENSE.pdf.
README.md ADDED
@@ -0,0 +1,285 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: pytorch
3
+ license: other
4
+ tags:
5
+ - android
6
+ pipeline_tag: image-segmentation
7
+
8
+ ---
9
+
10
+ ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/deeplab_xception/web-assets/model_demo.png)
11
+
12
+ # DeepLabXception: Optimized for Mobile Deployment
13
+ ## Deep Convolutional Neural Network model for semantic segmentation
14
+
15
+
16
+ DeepLabXception is a semantic segmentation model supporting multiple backbones like ResNet-101 and Xception, with flexible dataset compatibility including COCO, VOC, and Cityscapes.
17
+
18
+ This model is an implementation of DeepLabXception found [here](https://github.com/LikeLy-Journey/SegmenTron).
19
+
20
+
21
+ This repository provides scripts to run DeepLabXception on Qualcomm® devices.
22
+ More details on model performance across various devices, can be found
23
+ [here](https://aihub.qualcomm.com/models/deeplab_xception).
24
+
25
+
26
+
27
+ ### Model Details
28
+
29
+ - **Model Type:** Model_use_case.semantic_segmentation
30
+ - **Model Stats:**
31
+ - Model checkpoint: COCO_WITH_VOC_LABELS_V1
32
+ - Input resolution: 480x520
33
+ - Number of output classes: 21
34
+ - Number of parameters: 41.26M
35
+ - Model size (float): 158 MB
36
+
37
+ | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
+ |---|---|---|---|---|---|---|---|---|
39
+ | DeepLabXception | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 120.892 ms | 0 - 171 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
40
+ | DeepLabXception | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 112.829 ms | 0 - 76 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
41
+ | DeepLabXception | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 40.778 ms | 0 - 172 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
42
+ | DeepLabXception | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 52.14 ms | 0 - 79 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
43
+ | DeepLabXception | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 24.657 ms | 0 - 26 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
44
+ | DeepLabXception | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 21.564 ms | 3 - 33 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
45
+ | DeepLabXception | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 34.888 ms | 0 - 170 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
46
+ | DeepLabXception | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 30.527 ms | 0 - 76 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
47
+ | DeepLabXception | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 120.892 ms | 0 - 171 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
48
+ | DeepLabXception | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 112.829 ms | 0 - 76 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
49
+ | DeepLabXception | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 24.881 ms | 0 - 23 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
50
+ | DeepLabXception | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 21.888 ms | 3 - 32 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
51
+ | DeepLabXception | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 45.182 ms | 0 - 167 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
52
+ | DeepLabXception | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 39.099 ms | 0 - 81 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
53
+ | DeepLabXception | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 24.559 ms | 0 - 26 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
54
+ | DeepLabXception | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 21.801 ms | 2 - 32 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
55
+ | DeepLabXception | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 34.888 ms | 0 - 170 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
56
+ | DeepLabXception | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 30.527 ms | 0 - 76 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
57
+ | DeepLabXception | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 24.776 ms | 0 - 22 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
58
+ | DeepLabXception | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 21.808 ms | 3 - 33 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
59
+ | DeepLabXception | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 21.94 ms | 0 - 115 MB | NPU | [DeepLabXception.onnx.zip](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.onnx.zip) |
60
+ | DeepLabXception | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 18.293 ms | 0 - 196 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
61
+ | DeepLabXception | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 15.963 ms | 3 - 110 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
62
+ | DeepLabXception | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 16.518 ms | 3 - 97 MB | NPU | [DeepLabXception.onnx.zip](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.onnx.zip) |
63
+ | DeepLabXception | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 16.852 ms | 0 - 173 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.tflite) |
64
+ | DeepLabXception | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 14.019 ms | 3 - 85 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
65
+ | DeepLabXception | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 14.651 ms | 3 - 70 MB | NPU | [DeepLabXception.onnx.zip](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.onnx.zip) |
66
+ | DeepLabXception | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 22.614 ms | 189 - 189 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.dlc) |
67
+ | DeepLabXception | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 23.402 ms | 85 - 85 MB | NPU | [DeepLabXception.onnx.zip](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception.onnx.zip) |
68
+ | DeepLabXception | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 20.477 ms | 0 - 114 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
69
+ | DeepLabXception | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 20.663 ms | 1 - 142 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
70
+ | DeepLabXception | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 9.723 ms | 0 - 129 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
71
+ | DeepLabXception | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 15.459 ms | 1 - 146 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
72
+ | DeepLabXception | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 7.536 ms | 0 - 21 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
73
+ | DeepLabXception | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 8.032 ms | 1 - 31 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
74
+ | DeepLabXception | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 7.983 ms | 0 - 112 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
75
+ | DeepLabXception | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 8.241 ms | 1 - 138 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
76
+ | DeepLabXception | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 37.192 ms | 0 - 119 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
77
+ | DeepLabXception | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 55.477 ms | 1 - 136 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
78
+ | DeepLabXception | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 222.004 ms | 21 - 37 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
79
+ | DeepLabXception | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 20.477 ms | 0 - 114 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
80
+ | DeepLabXception | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 20.663 ms | 1 - 142 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
81
+ | DeepLabXception | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 7.568 ms | 0 - 18 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
82
+ | DeepLabXception | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 8.047 ms | 0 - 27 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
83
+ | DeepLabXception | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 12.857 ms | 0 - 113 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
84
+ | DeepLabXception | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 12.906 ms | 1 - 131 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
85
+ | DeepLabXception | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 7.536 ms | 0 - 22 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
86
+ | DeepLabXception | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 8.041 ms | 1 - 26 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
87
+ | DeepLabXception | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 7.983 ms | 0 - 112 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
88
+ | DeepLabXception | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 8.241 ms | 1 - 138 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
89
+ | DeepLabXception | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 7.525 ms | 0 - 22 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
90
+ | DeepLabXception | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 8.031 ms | 0 - 31 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
91
+ | DeepLabXception | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 5.41 ms | 0 - 139 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
92
+ | DeepLabXception | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 5.673 ms | 0 - 161 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
93
+ | DeepLabXception | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 5.039 ms | 0 - 115 MB | NPU | [DeepLabXception.tflite](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.tflite) |
94
+ | DeepLabXception | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 4.3 ms | 1 - 128 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
95
+ | DeepLabXception | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 8.708 ms | 131 - 131 MB | NPU | [DeepLabXception.dlc](https://huggingface.co/qualcomm/DeepLabXception/blob/main/DeepLabXception_w8a8.dlc) |
96
+
97
+
98
+
99
+
100
+ ## Installation
101
+
102
+
103
+ Install the package via pip:
104
+ ```bash
105
+ pip install qai-hub-models
106
+ ```
107
+
108
+
109
+ ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
110
+
111
+ Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
112
+ Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
113
+
114
+ With this API token, you can configure your client to run models on the cloud
115
+ hosted devices.
116
+ ```bash
117
+ qai-hub configure --api_token API_TOKEN
118
+ ```
119
+ Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information.
120
+
121
+
122
+
123
+ ## Demo off target
124
+
125
+ The package contains a simple end-to-end demo that downloads pre-trained
126
+ weights and runs this model on a sample input.
127
+
128
+ ```bash
129
+ python -m qai_hub_models.models.deeplab_xception.demo
130
+ ```
131
+
132
+ The above demo runs a reference implementation of pre-processing, model
133
+ inference, and post processing.
134
+
135
+ **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
136
+ environment, please add the following to your cell (instead of the above).
137
+ ```
138
+ %run -m qai_hub_models.models.deeplab_xception.demo
139
+ ```
140
+
141
+
142
+ ### Run model on a cloud-hosted device
143
+
144
+ In addition to the demo, you can also run the model on a cloud-hosted Qualcomm®
145
+ device. This script does the following:
146
+ * Performance check on-device on a cloud-hosted device
147
+ * Downloads compiled assets that can be deployed on-device for Android.
148
+ * Accuracy check between PyTorch and on-device outputs.
149
+
150
+ ```bash
151
+ python -m qai_hub_models.models.deeplab_xception.export
152
+ ```
153
+
154
+
155
+
156
+ ## How does this work?
157
+
158
+ This [export script](https://aihub.qualcomm.com/models/deeplab_xception/qai_hub_models/models/DeepLabXception/export.py)
159
+ leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) to optimize, validate, and deploy this model
160
+ on-device. Lets go through each step below in detail:
161
+
162
+ Step 1: **Compile model for on-device deployment**
163
+
164
+ To compile a PyTorch model for on-device deployment, we first trace the model
165
+ in memory using the `jit.trace` and then call the `submit_compile_job` API.
166
+
167
+ ```python
168
+ import torch
169
+
170
+ import qai_hub as hub
171
+ from qai_hub_models.models.deeplab_xception import Model
172
+
173
+ # Load the model
174
+ torch_model = Model.from_pretrained()
175
+
176
+ # Device
177
+ device = hub.Device("Samsung Galaxy S24")
178
+
179
+ # Trace model
180
+ input_shape = torch_model.get_input_spec()
181
+ sample_inputs = torch_model.sample_inputs()
182
+
183
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
184
+
185
+ # Compile model on a specific device
186
+ compile_job = hub.submit_compile_job(
187
+ model=pt_model,
188
+ device=device,
189
+ input_specs=torch_model.get_input_spec(),
190
+ )
191
+
192
+ # Get target model to run on-device
193
+ target_model = compile_job.get_target_model()
194
+
195
+ ```
196
+
197
+
198
+ Step 2: **Performance profiling on cloud-hosted device**
199
+
200
+ After compiling models from step 1. Models can be profiled model on-device using the
201
+ `target_model`. Note that this scripts runs the model on a device automatically
202
+ provisioned in the cloud. Once the job is submitted, you can navigate to a
203
+ provided job URL to view a variety of on-device performance metrics.
204
+ ```python
205
+ profile_job = hub.submit_profile_job(
206
+ model=target_model,
207
+ device=device,
208
+ )
209
+
210
+ ```
211
+
212
+ Step 3: **Verify on-device accuracy**
213
+
214
+ To verify the accuracy of the model on-device, you can run on-device inference
215
+ on sample input data on the same cloud hosted device.
216
+ ```python
217
+ input_data = torch_model.sample_inputs()
218
+ inference_job = hub.submit_inference_job(
219
+ model=target_model,
220
+ device=device,
221
+ inputs=input_data,
222
+ )
223
+ on_device_output = inference_job.download_output_data()
224
+
225
+ ```
226
+ With the output of the model, you can compute like PSNR, relative errors or
227
+ spot check the output with expected output.
228
+
229
+ **Note**: This on-device profiling and inference requires access to Qualcomm®
230
+ AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
231
+
232
+
233
+
234
+ ## Run demo on a cloud-hosted device
235
+
236
+ You can also run the demo on-device.
237
+
238
+ ```bash
239
+ python -m qai_hub_models.models.deeplab_xception.demo --eval-mode on-device
240
+ ```
241
+
242
+ **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
243
+ environment, please add the following to your cell (instead of the above).
244
+ ```
245
+ %run -m qai_hub_models.models.deeplab_xception.demo -- --eval-mode on-device
246
+ ```
247
+
248
+
249
+ ## Deploying compiled model to Android
250
+
251
+
252
+ The models can be deployed using multiple runtimes:
253
+ - TensorFlow Lite (`.tflite` export): [This
254
+ tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a
255
+ guide to deploy the .tflite model in an Android application.
256
+
257
+
258
+ - QNN (`.so` export ): This [sample
259
+ app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html)
260
+ provides instructions on how to use the `.so` shared library in an Android application.
261
+
262
+
263
+ ## View on Qualcomm® AI Hub
264
+ Get more details on DeepLabXception's performance across various devices [here](https://aihub.qualcomm.com/models/deeplab_xception).
265
+ Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
266
+
267
+
268
+ ## License
269
+ * The license for the original implementation of DeepLabXception can be found
270
+ [here](https://github.com/pytorch/vision/blob/main/LICENSE).
271
+ * The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf)
272
+
273
+
274
+
275
+ ## References
276
+ * [Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation](https://arxiv.org/abs/1802.02611)
277
+ * [Source Model Implementation](https://github.com/LikeLy-Journey/SegmenTron)
278
+
279
+
280
+
281
+ ## Community
282
+ * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
283
+ * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).
284
+
285
+
precompiled/qualcomm-qcs6490-proxy/DeepLabXception_w8a8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bcabbacd86f26e4e7f024fdf8e9608f6ba86d81c5e3b904149c4e269e0744b32
3
+ size 57774080
precompiled/qualcomm-qcs6490-proxy/sdk_versions.yml ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ sdk_versions:
2
+ qnn_context_binary:
3
+ qairt: 2.37.0.250724175447_124859
precompiled/qualcomm-snapdragon-x-elite/DeepLabXception.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc09fbdf25b8af1f598b7e8cdef76daf6aa0da0e5c00362f62d97d8bcb40974b
3
+ size 91877376
precompiled/qualcomm-snapdragon-x-elite/DeepLabXception.onnx.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f21f306251d4739f5fcd18ae820af5bb63cf5c7ace45b216b0e7d3354a7bfe8
3
+ size 77832392
precompiled/qualcomm-snapdragon-x-elite/DeepLabXception_w8a8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9225f9fbb54859c61020615b623e2376d3889aa943cd66dbda3d36f7ca7eb738
3
+ size 46477312
precompiled/qualcomm-snapdragon-x-elite/DeepLabXception_w8a8.onnx.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:442834c707c6093ee2126e2231881cd63e4a4b4bbb85c7b1a98162533b2cb9b5
3
+ size 37026952
precompiled/qualcomm-snapdragon-x-elite/sdk_versions.yml ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ sdk_versions:
2
+ qnn_context_binary:
3
+ qairt: 2.37.0.250724175447_124859
4
+ precompiled_qnn_onnx:
5
+ qairt: 2.33.2.250410134701_117956
sdk_versions.yml ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ sdk_versions:
2
+ tflite:
3
+ qairt: 2.37.0.250724175447_124859
4
+ tflite: 2.17.0
5
+ qnn_dlc:
6
+ qairt: 2.37.0.250724175447_124859
7
+ onnx:
8
+ qairt: 2.33.2.250410134701_117956
9
+ onnx_runtime: 1.22.0