Hugging Face's logo

qualcomm
/
Detectron2-Detection

Object Detection
PyTorch
LiteRT
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
daa6fcb
·
verified ·
1 Parent(s): 08b9293

v0.44.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.44.0 for changelog.

Files changed (16) hide show
  1. .gitattributes +6 -0
  2. Detectron2-Detection_Detectron2ProposalGenerator_float.dlc +3 -0
  3. Detectron2-Detection_Detectron2ProposalGenerator_float.tflite +3 -0
  4. Detectron2-Detection_Detectron2ProposalGenerator_w8a16.dlc +3 -0
  5. Detectron2-Detection_Detectron2ProposalGenerator_w8a16.onnx.zip +3 -0
  6. Detectron2-Detection_Detectron2ProposalGenerator_w8a8.dlc +3 -0
  7. Detectron2-Detection_Detectron2ProposalGenerator_w8a8.onnx.zip +3 -0
  8. Detectron2-Detection_Detectron2ROIHead_float.dlc +3 -0
  9. Detectron2-Detection_Detectron2ROIHead_float.tflite +3 -0
  10. Detectron2-Detection_Detectron2ROIHead_w8a16.dlc +3 -0
  11. Detectron2-Detection_Detectron2ROIHead_w8a16.onnx.zip +3 -0
  12. Detectron2-Detection_Detectron2ROIHead_w8a8.dlc +3 -0
  13. Detectron2-Detection_Detectron2ROIHead_w8a8.onnx.zip +3 -0
  14. LICENSE +1 -0
  15. README.md +353 -0
  16. tool-versions.yaml +4 -0
.gitattributes CHANGED
@@ -33,3 +33,9 @@ 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
+ Detectron2-Detection_Detectron2ProposalGenerator_float.dlc filter=lfs diff=lfs merge=lfs -text
37
+ Detectron2-Detection_Detectron2ProposalGenerator_w8a16.dlc filter=lfs diff=lfs merge=lfs -text
38
+ Detectron2-Detection_Detectron2ProposalGenerator_w8a8.dlc filter=lfs diff=lfs merge=lfs -text
39
+ Detectron2-Detection_Detectron2ROIHead_float.dlc filter=lfs diff=lfs merge=lfs -text
40
+ Detectron2-Detection_Detectron2ROIHead_w8a16.dlc filter=lfs diff=lfs merge=lfs -text
41
+ Detectron2-Detection_Detectron2ROIHead_w8a8.dlc filter=lfs diff=lfs merge=lfs -text
Detectron2-Detection_Detectron2ProposalGenerator_float.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:731b50dee0bad4bbd56d7cbd3584634a8c976e0a864919b90a60275e83b5fe11
3
+ size 72910388
Detectron2-Detection_Detectron2ProposalGenerator_float.tflite ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9940797d7ffdb96332fbd1f50a9af650b40383a0dd539b469b0b0af064b27b43
3
+ size 72802024
Detectron2-Detection_Detectron2ProposalGenerator_w8a16.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b06afd371f1f60ca4c6e68a8bfba0fff2a1c46a283af3e043abddc8e31e7833
3
+ size 19455796
Detectron2-Detection_Detectron2ProposalGenerator_w8a16.onnx.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8abe1e76827b33465edf977eb1178bee434c792f947088b008caee8cebc95f1c
3
+ size 31298196
Detectron2-Detection_Detectron2ProposalGenerator_w8a8.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40216fd9ade37cf232737f4258377cf4620fa54f4f23eade6d0e6bacedf3c365
3
+ size 19305788
Detectron2-Detection_Detectron2ProposalGenerator_w8a8.onnx.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e775ec5f921779872539096d1b97e658f23a3ec19876d73903c0f64f4cb52b3d
3
+ size 31380433
Detectron2-Detection_Detectron2ROIHead_float.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:132954b3f5840a0407da8f6d17a4c7e8c8473dc3d8f5f8f9e6b39233358de116
3
+ size 63197524
Detectron2-Detection_Detectron2ROIHead_float.tflite ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd9cb74ca9a03874b6e486030009a84d7963d2379af01110db24a26d45706e6c
3
+ size 63293948
Detectron2-Detection_Detectron2ROIHead_w8a16.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2344d092b1b7273d3ea604a612d9b966c0e03ca6d68c85e8966ab6cb4d63ff94
3
+ size 16544580
Detectron2-Detection_Detectron2ROIHead_w8a16.onnx.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7adbff4df7b207d285a6cf73a73ac7baad8fbadcdecd322d2a9a2465d5b1a0b
3
+ size 26485841
Detectron2-Detection_Detectron2ROIHead_w8a8.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:288b074d3cc30eae192b64284c08ccf2039dd1084e1224ce4d6d632c97bb6677
3
+ size 16544372
Detectron2-Detection_Detectron2ROIHead_w8a8.onnx.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37806372ee10f8e207ed9fa56bc0f11fa0de1d4fa98109abf22d061e957df92a
3
+ size 26587837
LICENSE ADDED
@@ -0,0 +1 @@
 
 
1
+ The license of the original trained model can be found at https://github.com/facebookresearch/detectron2/blob/main/LICENSE.
README.md ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: pytorch
3
+ license: other
4
+ tags:
5
+ - android
6
+ pipeline_tag: object-detection
7
+
8
+ ---
9
+
10
+ ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/detectron2_detection/web-assets/model_demo.png)
11
+
12
+ # Detectron2-Detection: Optimized for Mobile Deployment
13
+ ## A next-generation library for object detection
14
+
15
+
16
+ Detectron2-Detection is a machine learning model that can detect objects (trained on COCO dataset).
17
+
18
+ This model is an implementation of Detectron2-Detection found [here](https://github.com/facebookresearch/detectron2/).
19
+
20
+
21
+ This repository provides scripts to run Detectron2-Detection on Qualcomm® devices.
22
+ More details on model performance across various devices, can be found
23
+ [here](https://aihub.qualcomm.com/models/detectron2_detection).
24
+
25
+
26
+
27
+ ### Model Details
28
+
29
+ - **Model Type:** Model_use_case.object_detection
30
+ - **Model Stats:**
31
+ - Model checkpoint: faster_rcnn_R_50_C4_1x
32
+ - Input resolution: 800x800
33
+ - Number of parameters: 33.8M
34
+ - Model size (float): 129 MB
35
+
36
+ | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
37
+ |---|---|---|---|---|---|---|---|---|
38
+ | Detectron2ProposalGenerator | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 187.314 ms | 0 - 426 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
39
+ | Detectron2ProposalGenerator | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 88.096 ms | 7 - 469 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
40
+ | Detectron2ProposalGenerator | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 56.493 ms | 7 - 10 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
41
+ | Detectron2ProposalGenerator | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 69.755 ms | 1 - 428 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
42
+ | Detectron2ProposalGenerator | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 187.314 ms | 0 - 426 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
43
+ | Detectron2ProposalGenerator | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 56.53 ms | 7 - 10 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
44
+ | Detectron2ProposalGenerator | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 62.02 ms | 11 - 413 MB | NPU | [Detectron2-Detection.tflite](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.tflite) |
45
+ | Detectron2ProposalGenerator | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 79.497 ms | 0 - 393 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
46
+ | Detectron2ProposalGenerator | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 55.695 ms | 9 - 12 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
47
+ | Detectron2ProposalGenerator | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 69.755 ms | 1 - 428 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
48
+ | Detectron2ProposalGenerator | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 38.672 ms | 7 - 526 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
49
+ | Detectron2ProposalGenerator | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 23.367 ms | 9 - 438 MB | NPU | [Detectron2-Detection.tflite](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.tflite) |
50
+ | Detectron2ProposalGenerator | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 28.699 ms | 7 - 404 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
51
+ | Detectron2ProposalGenerator | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 16.65 ms | 10 - 403 MB | NPU | [Detectron2-Detection.tflite](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.tflite) |
52
+ | Detectron2ProposalGenerator | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 23.922 ms | 7 - 413 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
53
+ | Detectron2ProposalGenerator | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 56.926 ms | 7 - 7 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
54
+ | Detectron2ROIHead | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 330.557 ms | 4 - 1333 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
55
+ | Detectron2ROIHead | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 100.504 ms | 10 - 415 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
56
+ | Detectron2ROIHead | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 64.389 ms | 10 - 13 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
57
+ | Detectron2ROIHead | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 445.931 ms | 1 - 1330 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
58
+ | Detectron2ROIHead | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 330.557 ms | 4 - 1333 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
59
+ | Detectron2ROIHead | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 64.616 ms | 10 - 13 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
60
+ | Detectron2ROIHead | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 761.98 ms | 0 - 1533 MB | CPU | [Detectron2-Detection.tflite](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.tflite) |
61
+ | Detectron2ROIHead | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 89.6 ms | 0 - 358 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
62
+ | Detectron2ROIHead | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 64.938 ms | 10 - 17 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
63
+ | Detectron2ROIHead | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 445.931 ms | 1 - 1330 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
64
+ | Detectron2ROIHead | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 48.169 ms | 10 - 1932 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
65
+ | Detectron2ROIHead | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 2229.127 ms | 0 - 1721 MB | NPU | [Detectron2-Detection.tflite](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.tflite) |
66
+ | Detectron2ROIHead | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 40.828 ms | 10 - 1336 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
67
+ | Detectron2ROIHead | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 2001.288 ms | 0 - 1886 MB | NPU | [Detectron2-Detection.tflite](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.tflite) |
68
+ | Detectron2ROIHead | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 26.899 ms | 10 - 1414 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
69
+ | Detectron2ROIHead | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 64.294 ms | 10 - 10 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection.dlc) |
70
+ | Detectron2ProposalGenerator | w8a16 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | QNN_DLC | 355.555 ms | 4 - 512 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
71
+ | Detectron2ProposalGenerator | w8a16 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | QNN_DLC | 153.72 ms | 4 - 14 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
72
+ | Detectron2ProposalGenerator | w8a16 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | ONNX | 2995.84 ms | 189 - 197 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
73
+ | Detectron2ProposalGenerator | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 113.986 ms | 2 - 391 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
74
+ | Detectron2ProposalGenerator | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 69.449 ms | 4 - 515 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
75
+ | Detectron2ProposalGenerator | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 64.245 ms | 4 - 7 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
76
+ | Detectron2ProposalGenerator | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 19.569 ms | 15 - 19 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
77
+ | Detectron2ProposalGenerator | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 64.402 ms | 1 - 436 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
78
+ | Detectron2ProposalGenerator | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 113.986 ms | 2 - 391 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
79
+ | Detectron2ProposalGenerator | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 63.656 ms | 4 - 6 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
80
+ | Detectron2ProposalGenerator | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 74.127 ms | 0 - 434 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
81
+ | Detectron2ProposalGenerator | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 64.018 ms | 5 - 8 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
82
+ | Detectron2ProposalGenerator | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 64.402 ms | 1 - 436 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
83
+ | Detectron2ProposalGenerator | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 46.522 ms | 4 - 492 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
84
+ | Detectron2ProposalGenerator | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 13.528 ms | 17 - 409 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
85
+ | Detectron2ProposalGenerator | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 26.781 ms | 4 - 391 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
86
+ | Detectron2ProposalGenerator | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 10.64 ms | 17 - 310 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
87
+ | Detectron2ProposalGenerator | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 54.328 ms | 4 - 445 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
88
+ | Detectron2ProposalGenerator | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | ONNX | 1394.375 ms | 173 - 189 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
89
+ | Detectron2ProposalGenerator | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 22.864 ms | 4 - 419 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
90
+ | Detectron2ProposalGenerator | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 8.655 ms | 20 - 385 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
91
+ | Detectron2ProposalGenerator | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 66.822 ms | 4 - 4 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
92
+ | Detectron2ProposalGenerator | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 20.186 ms | 33 - 33 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
93
+ | Detectron2ROIHead | w8a16 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | QNN_DLC | 497.166 ms | 5 - 473 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
94
+ | Detectron2ROIHead | w8a16 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | QNN_DLC | 130.402 ms | 1 - 7 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
95
+ | Detectron2ROIHead | w8a16 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | ONNX | 8643.155 ms | 3039 - 3054 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
96
+ | Detectron2ROIHead | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 101.309 ms | 4 - 364 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
97
+ | Detectron2ROIHead | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 57.782 ms | 5 - 428 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
98
+ | Detectron2ROIHead | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 35.18 ms | 5 - 7 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
99
+ | Detectron2ROIHead | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 303.607 ms | 1014 - 1016 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
100
+ | Detectron2ROIHead | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 35.311 ms | 0 - 361 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
101
+ | Detectron2ROIHead | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 101.309 ms | 4 - 364 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
102
+ | Detectron2ROIHead | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 35.101 ms | 5 - 7 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
103
+ | Detectron2ROIHead | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 49.802 ms | 0 - 362 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
104
+ | Detectron2ROIHead | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 35.002 ms | 5 - 8 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
105
+ | Detectron2ROIHead | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 35.311 ms | 0 - 361 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
106
+ | Detectron2ROIHead | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 25.1 ms | 5 - 427 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
107
+ | Detectron2ROIHead | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 262.813 ms | 1008 - 1435 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
108
+ | Detectron2ROIHead | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 18.557 ms | 5 - 391 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
109
+ | Detectron2ROIHead | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 239.171 ms | 1001 - 1382 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
110
+ | Detectron2ROIHead | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 47.939 ms | 5 - 420 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
111
+ | Detectron2ROIHead | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | ONNX | 3592.796 ms | 3033 - 3048 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
112
+ | Detectron2ROIHead | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 14.274 ms | 5 - 393 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
113
+ | Detectron2ROIHead | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 188.861 ms | 1014 - 1399 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
114
+ | Detectron2ROIHead | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 36.239 ms | 5 - 5 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.dlc) |
115
+ | Detectron2ROIHead | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 256.69 ms | 1059 - 1059 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a16.onnx.zip) |
116
+ | Detectron2ProposalGenerator | w8a8 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | QNN_DLC | 186.888 ms | 2 - 377 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
117
+ | Detectron2ProposalGenerator | w8a8 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | ONNX | 449.325 ms | 121 - 136 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
118
+ | Detectron2ProposalGenerator | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | QNN_DLC | 63.304 ms | 2 - 8 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
119
+ | Detectron2ProposalGenerator | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | ONNX | 447.641 ms | 141 - 175 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
120
+ | Detectron2ProposalGenerator | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 72.247 ms | 2 - 371 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
121
+ | Detectron2ProposalGenerator | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 47.461 ms | 2 - 426 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
122
+ | Detectron2ProposalGenerator | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 40.044 ms | 2 - 4 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
123
+ | Detectron2ProposalGenerator | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 12.406 ms | 8 - 10 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
124
+ | Detectron2ProposalGenerator | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 183.946 ms | 1 - 371 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
125
+ | Detectron2ProposalGenerator | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 72.247 ms | 2 - 371 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
126
+ | Detectron2ProposalGenerator | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 40.081 ms | 2 - 4 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
127
+ | Detectron2ProposalGenerator | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 45.448 ms | 2 - 375 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
128
+ | Detectron2ProposalGenerator | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 40.117 ms | 2 - 5 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
129
+ | Detectron2ProposalGenerator | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 183.946 ms | 1 - 371 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
130
+ | Detectron2ProposalGenerator | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 26.892 ms | 2 - 436 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
131
+ | Detectron2ProposalGenerator | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 8.924 ms | 10 - 327 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
132
+ | Detectron2ProposalGenerator | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 22.806 ms | 2 - 358 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
133
+ | Detectron2ProposalGenerator | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 7.066 ms | 8 - 256 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
134
+ | Detectron2ProposalGenerator | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 29.502 ms | 2 - 379 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
135
+ | Detectron2ProposalGenerator | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | ONNX | 357.966 ms | 133 - 148 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
136
+ | Detectron2ProposalGenerator | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 23.106 ms | 2 - 367 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
137
+ | Detectron2ProposalGenerator | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 5.733 ms | 8 - 258 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
138
+ | Detectron2ProposalGenerator | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 42.189 ms | 2 - 2 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
139
+ | Detectron2ProposalGenerator | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 12.51 ms | 20 - 20 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
140
+ | Detectron2ROIHead | w8a8 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | QNN_DLC | 278.301 ms | 3 - 372 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
141
+ | Detectron2ROIHead | w8a8 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | ONNX | 1072.446 ms | 2473 - 2487 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
142
+ | Detectron2ROIHead | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | QNN_DLC | 80.145 ms | 4 - 8 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
143
+ | Detectron2ROIHead | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | ONNX | 1721.984 ms | 2531 - 2545 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
144
+ | Detectron2ROIHead | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 53.145 ms | 2 - 300 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
145
+ | Detectron2ROIHead | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 40.959 ms | 3 - 347 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
146
+ | Detectron2ROIHead | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 19.269 ms | 2 - 5 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
147
+ | Detectron2ROIHead | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 112.829 ms | 507 - 510 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
148
+ | Detectron2ROIHead | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 19.861 ms | 2 - 302 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
149
+ | Detectron2ROIHead | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 53.145 ms | 2 - 300 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
150
+ | Detectron2ROIHead | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 19.307 ms | 2 - 4 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
151
+ | Detectron2ROIHead | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 29.261 ms | 0 - 297 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
152
+ | Detectron2ROIHead | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 19.239 ms | 2 - 5 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
153
+ | Detectron2ROIHead | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 19.861 ms | 2 - 302 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
154
+ | Detectron2ROIHead | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 13.498 ms | 2 - 354 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
155
+ | Detectron2ROIHead | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 93.183 ms | 507 - 843 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
156
+ | Detectron2ROIHead | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 10.295 ms | 2 - 310 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
157
+ | Detectron2ROIHead | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 88.013 ms | 467 - 757 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
158
+ | Detectron2ROIHead | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 29.544 ms | 2 - 324 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
159
+ | Detectron2ROIHead | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | ONNX | 983.931 ms | 2509 - 2525 MB | CPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
160
+ | Detectron2ROIHead | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 8.368 ms | 2 - 310 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
161
+ | Detectron2ROIHead | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 68.356 ms | 509 - 802 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
162
+ | Detectron2ROIHead | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 19.97 ms | 2 - 2 MB | NPU | [Detectron2-Detection.dlc](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.dlc) |
163
+ | Detectron2ROIHead | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 100.047 ms | 530 - 530 MB | NPU | [Detectron2-Detection.onnx.zip](https://huggingface.co/qualcomm/Detectron2-Detection/blob/main/Detectron2-Detection_w8a8.onnx.zip) |
164
+
165
+
166
+
167
+
168
+ ## Installation
169
+
170
+
171
+ Install the package via pip:
172
+ ```bash
173
+ # NOTE: 3.10 <= PYTHON_VERSION < 3.14 is supported.
174
+ pip install wheel==0.45.1 "torch>=2.1,<2.9.0" "setuptools>=77.0.3"
175
+ pip install "qai-hub-models[detectron2-detection]" git+https://github.com/facebookresearch/detectron2.git@d38d716 --no-build-isolation
176
+ ```
177
+
178
+
179
+ ## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
180
+
181
+ Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.com/) with your
182
+ Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
183
+
184
+ With this API token, you can configure your client to run models on the cloud
185
+ hosted devices.
186
+ ```bash
187
+ qai-hub configure --api_token API_TOKEN
188
+ ```
189
+ Navigate to [docs](https://workbench.aihub.qualcomm.com/docs/) for more information.
190
+
191
+
192
+
193
+ ## Demo off target
194
+
195
+ The package contains a simple end-to-end demo that downloads pre-trained
196
+ weights and runs this model on a sample input.
197
+
198
+ ```bash
199
+ python -m qai_hub_models.models.detectron2_detection.demo
200
+ ```
201
+
202
+ The above demo runs a reference implementation of pre-processing, model
203
+ inference, and post processing.
204
+
205
+ **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
206
+ environment, please add the following to your cell (instead of the above).
207
+ ```
208
+ %run -m qai_hub_models.models.detectron2_detection.demo
209
+ ```
210
+
211
+
212
+ ### Run model on a cloud-hosted device
213
+
214
+ In addition to the demo, you can also run the model on a cloud-hosted Qualcomm®
215
+ device. This script does the following:
216
+ * Performance check on-device on a cloud-hosted device
217
+ * Downloads compiled assets that can be deployed on-device for Android.
218
+ * Accuracy check between PyTorch and on-device outputs.
219
+
220
+ ```bash
221
+ python -m qai_hub_models.models.detectron2_detection.export
222
+ ```
223
+
224
+
225
+
226
+ ## How does this work?
227
+
228
+ This [export script](https://aihub.qualcomm.com/models/detectron2_detection/qai_hub_models/models/Detectron2-Detection/export.py)
229
+ leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) to optimize, validate, and deploy this model
230
+ on-device. Lets go through each step below in detail:
231
+
232
+ Step 1: **Compile model for on-device deployment**
233
+
234
+ To compile a PyTorch model for on-device deployment, we first trace the model
235
+ in memory using the `jit.trace` and then call the `submit_compile_job` API.
236
+
237
+ ```python
238
+ import torch
239
+
240
+ import qai_hub as hub
241
+ from qai_hub_models.models.detectron2_detection import Model
242
+
243
+ # Load the model
244
+ torch_model = Model.from_pretrained()
245
+
246
+ # Device
247
+ device = hub.Device("Samsung Galaxy S25")
248
+
249
+ # Trace model
250
+ input_shape = torch_model.get_input_spec()
251
+ sample_inputs = torch_model.sample_inputs()
252
+
253
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
254
+
255
+ # Compile model on a specific device
256
+ compile_job = hub.submit_compile_job(
257
+ model=pt_model,
258
+ device=device,
259
+ input_specs=torch_model.get_input_spec(),
260
+ )
261
+
262
+ # Get target model to run on-device
263
+ target_model = compile_job.get_target_model()
264
+
265
+ ```
266
+
267
+
268
+ Step 2: **Performance profiling on cloud-hosted device**
269
+
270
+ After compiling models from step 1. Models can be profiled model on-device using the
271
+ `target_model`. Note that this scripts runs the model on a device automatically
272
+ provisioned in the cloud. Once the job is submitted, you can navigate to a
273
+ provided job URL to view a variety of on-device performance metrics.
274
+ ```python
275
+ profile_job = hub.submit_profile_job(
276
+ model=target_model,
277
+ device=device,
278
+ )
279
+
280
+ ```
281
+
282
+ Step 3: **Verify on-device accuracy**
283
+
284
+ To verify the accuracy of the model on-device, you can run on-device inference
285
+ on sample input data on the same cloud hosted device.
286
+ ```python
287
+ input_data = torch_model.sample_inputs()
288
+ inference_job = hub.submit_inference_job(
289
+ model=target_model,
290
+ device=device,
291
+ inputs=input_data,
292
+ )
293
+ on_device_output = inference_job.download_output_data()
294
+
295
+ ```
296
+ With the output of the model, you can compute like PSNR, relative errors or
297
+ spot check the output with expected output.
298
+
299
+ **Note**: This on-device profiling and inference requires access to Qualcomm®
300
+ AI Hub Workbench. [Sign up for access](https://myaccount.qualcomm.com/signup).
301
+
302
+
303
+
304
+ ## Run demo on a cloud-hosted device
305
+
306
+ You can also run the demo on-device.
307
+
308
+ ```bash
309
+ python -m qai_hub_models.models.detectron2_detection.demo --eval-mode on-device
310
+ ```
311
+
312
+ **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
313
+ environment, please add the following to your cell (instead of the above).
314
+ ```
315
+ %run -m qai_hub_models.models.detectron2_detection.demo -- --eval-mode on-device
316
+ ```
317
+
318
+
319
+ ## Deploying compiled model to Android
320
+
321
+
322
+ The models can be deployed using multiple runtimes:
323
+ - TensorFlow Lite (`.tflite` export): [This
324
+ tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a
325
+ guide to deploy the .tflite model in an Android application.
326
+
327
+
328
+ - QNN (`.so` export ): This [sample
329
+ app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html)
330
+ provides instructions on how to use the `.so` shared library in an Android application.
331
+
332
+
333
+ ## View on Qualcomm® AI Hub
334
+ Get more details on Detectron2-Detection's performance across various devices [here](https://aihub.qualcomm.com/models/detectron2_detection).
335
+ Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
336
+
337
+
338
+ ## License
339
+ * The license for the original implementation of Detectron2-Detection can be found
340
+ [here](https://github.com/facebookresearch/detectron2/blob/main/LICENSE).
341
+
342
+
343
+
344
+ ## References
345
+ * [Source Model Implementation](https://github.com/facebookresearch/detectron2/)
346
+
347
+
348
+
349
+ ## Community
350
+ * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
351
+ * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).
352
+
353
+
tool-versions.yaml ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ tool_versions:
2
+ onnx:
3
+ qairt: 2.37.1.250807093845_124904
4
+ onnx_runtime: 1.23.0