qaihm-bot commited on
Commit
783afe2
·
verified ·
1 Parent(s): 4242cab

See https://github.com/quic/ai-hub-models/releases/v0.42.0 for changelog.

GoogLeNet_float.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ff355850aa87a6646b6f5280c6804989eeda6de4c0c0634db5860d23111c974f
3
- size 26640996
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f719ffc546f9d551e8ea01f3885a8a6804dd2f1017bb8cd3d58910e1bef2faa
3
+ size 26641100
GoogLeNet_float.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a5a849d92fe516b73546083793fee491b7b66071856f8550991d93ee6896d69c
3
  size 24669348
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed20e1d28a9eacac2f3469870339d59495487b85c41a913f143cd92b8c368a4f
3
  size 24669348
GoogLeNet_w8a8.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9802cfc23218ac8ad616d727f743f3e03385e2ba82c0125403447d1d63b30bf3
3
- size 7228204
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1688cf9f824af0159c50879ca9168afd775e6dea499ff853899fd9c2c96c3987
3
+ size 7228308
GoogLeNet_w8a8.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6784c5b2b6ec455774612941ad27986793d86f2bdd5af0fff1b81de137830332
3
  size 11844712
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3332bc98697a14abeb277666c33a36ceeaae817e7f7b6efac9ac9a60c9aa9a0d
3
  size 11844712
GoogLeNet_w8a8.tflite CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:56dc4247bcf8457c13ad6866b786391822396e1ad5a13b23b67f5f7d1a8f6702
3
  size 6857952
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f031fd900fd57c7404344a15e51c2fde3fb61d29c93bff1937adc28e2061d810
3
  size 6857952
README.md CHANGED
@@ -36,71 +36,74 @@ More details on model performance across various devices, can be found
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
- | GoogLeNet | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 5.114 ms | 0 - 25 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
40
- | GoogLeNet | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 5.056 ms | 1 - 21 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
41
- | GoogLeNet | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.395 ms | 0 - 42 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
42
- | GoogLeNet | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.767 ms | 1 - 33 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
43
- | GoogLeNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 0.934 ms | 0 - 95 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
44
- | GoogLeNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 0.832 ms | 0 - 22 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
45
- | GoogLeNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 1.081 ms | 0 - 34 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
46
- | GoogLeNet | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.66 ms | 0 - 25 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
47
- | GoogLeNet | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.567 ms | 1 - 21 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
48
- | GoogLeNet | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 5.114 ms | 0 - 25 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
49
- | GoogLeNet | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 5.056 ms | 1 - 21 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
50
- | GoogLeNet | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 0.906 ms | 0 - 96 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
51
- | GoogLeNet | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 0.834 ms | 1 - 4 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
52
- | GoogLeNet | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 1.87 ms | 0 - 32 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
53
- | GoogLeNet | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 1.795 ms | 1 - 25 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
54
- | GoogLeNet | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 0.934 ms | 0 - 95 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
55
- | GoogLeNet | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 0.835 ms | 0 - 43 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
56
- | GoogLeNet | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 1.66 ms | 0 - 25 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
57
- | GoogLeNet | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.567 ms | 1 - 21 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
58
- | GoogLeNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.612 ms | 0 - 42 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
59
- | GoogLeNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.589 ms | 1 - 30 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
60
- | GoogLeNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 0.677 ms | 0 - 33 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
61
- | GoogLeNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 0.489 ms | 0 - 33 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
62
- | GoogLeNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.458 ms | 1 - 28 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
63
- | GoogLeNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 0.571 ms | 0 - 26 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
64
- | GoogLeNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 0.411 ms | 0 - 33 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
65
- | GoogLeNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 0.38 ms | 1 - 27 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
66
- | GoogLeNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 0.527 ms | 1 - 27 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
67
- | GoogLeNet | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 0.998 ms | 33 - 33 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
68
- | GoogLeNet | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.053 ms | 13 - 13 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
69
- | GoogLeNet | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 0.866 ms | 0 - 20 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
70
- | GoogLeNet | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 0.799 ms | 0 - 21 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
71
- | GoogLeNet | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 0.34 ms | 0 - 34 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
72
- | GoogLeNet | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 0.427 ms | 0 - 36 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
73
- | GoogLeNet | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 0.27 ms | 0 - 39 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
74
- | GoogLeNet | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 0.243 ms | 0 - 40 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
75
- | GoogLeNet | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 0.524 ms | 0 - 48 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
76
- | GoogLeNet | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 0.454 ms | 0 - 21 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
77
- | GoogLeNet | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 0.422 ms | 0 - 21 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
78
- | GoogLeNet | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 19.761 ms | 1 - 24 MB | GPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
79
- | GoogLeNet | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 1.067 ms | 0 - 28 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
80
- | GoogLeNet | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 9.691 ms | 9 - 25 MB | CPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
81
- | GoogLeNet | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 6.245 ms | 0 - 11 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
82
- | GoogLeNet | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 8.455 ms | 9 - 18 MB | CPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
83
- | GoogLeNet | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 0.866 ms | 0 - 20 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
84
- | GoogLeNet | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 0.799 ms | 0 - 21 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
85
- | GoogLeNet | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 0.252 ms | 0 - 39 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
86
- | GoogLeNet | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 0.244 ms | 0 - 40 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
87
- | GoogLeNet | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 0.638 ms | 0 - 26 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
88
- | GoogLeNet | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 0.604 ms | 0 - 27 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
89
- | GoogLeNet | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 0.256 ms | 0 - 40 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
90
- | GoogLeNet | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 0.25 ms | 0 - 40 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
91
- | GoogLeNet | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 0.454 ms | 0 - 21 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
92
- | GoogLeNet | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 0.422 ms | 0 - 21 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
93
- | GoogLeNet | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.2 ms | 0 - 36 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
94
- | GoogLeNet | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.185 ms | 0 - 38 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
95
- | GoogLeNet | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 0.353 ms | 0 - 41 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
96
- | GoogLeNet | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 0.161 ms | 0 - 23 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
97
- | GoogLeNet | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.147 ms | 0 - 29 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
98
- | GoogLeNet | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 0.318 ms | 0 - 34 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
99
- | GoogLeNet | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 0.147 ms | 0 - 23 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
100
- | GoogLeNet | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 0.143 ms | 0 - 25 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
101
- | GoogLeNet | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 0.314 ms | 0 - 30 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
102
- | GoogLeNet | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 0.339 ms | 30 - 30 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
103
- | GoogLeNet | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 0.466 ms | 8 - 8 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
 
 
 
104
 
105
 
106
 
@@ -114,9 +117,9 @@ pip install qai-hub-models
114
  ```
115
 
116
 
117
- ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
118
 
119
- Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
120
  Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
121
 
122
  With this API token, you can configure your client to run models on the cloud
@@ -124,7 +127,7 @@ hosted devices.
124
  ```bash
125
  qai-hub configure --api_token API_TOKEN
126
  ```
127
- Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information.
128
 
129
 
130
 
@@ -235,7 +238,7 @@ With the output of the model, you can compute like PSNR, relative errors or
235
  spot check the output with expected output.
236
 
237
  **Note**: This on-device profiling and inference requires access to Qualcomm®
238
- AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
239
 
240
 
241
 
 
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
+ | GoogLeNet | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 5.155 ms | 0 - 28 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
40
+ | GoogLeNet | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 5.006 ms | 1 - 22 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
41
+ | GoogLeNet | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.4 ms | 0 - 43 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
42
+ | GoogLeNet | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.762 ms | 1 - 31 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
43
+ | GoogLeNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 0.937 ms | 0 - 92 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
44
+ | GoogLeNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 0.84 ms | 0 - 30 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
45
+ | GoogLeNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 1.104 ms | 0 - 42 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
46
+ | GoogLeNet | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 7.282 ms | 0 - 28 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
47
+ | GoogLeNet | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.561 ms | 1 - 22 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
48
+ | GoogLeNet | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 5.155 ms | 0 - 28 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
49
+ | GoogLeNet | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 5.006 ms | 1 - 22 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
50
+ | GoogLeNet | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 0.945 ms | 0 - 94 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
51
+ | GoogLeNet | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 0.838 ms | 0 - 42 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
52
+ | GoogLeNet | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 1.896 ms | 0 - 34 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
53
+ | GoogLeNet | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 1.789 ms | 1 - 27 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
54
+ | GoogLeNet | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 0.94 ms | 0 - 93 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
55
+ | GoogLeNet | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 0.841 ms | 0 - 47 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
56
+ | GoogLeNet | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 7.282 ms | 0 - 28 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
57
+ | GoogLeNet | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.561 ms | 1 - 22 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
58
+ | GoogLeNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.597 ms | 0 - 37 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
59
+ | GoogLeNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.577 ms | 1 - 33 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
60
+ | GoogLeNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 0.675 ms | 0 - 29 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
61
+ | GoogLeNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 0.491 ms | 0 - 34 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
62
+ | GoogLeNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.461 ms | 1 - 29 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
63
+ | GoogLeNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 0.573 ms | 0 - 26 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
64
+ | GoogLeNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 0.411 ms | 0 - 34 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.tflite) |
65
+ | GoogLeNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 0.38 ms | 0 - 27 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
66
+ | GoogLeNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 0.514 ms | 1 - 23 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
67
+ | GoogLeNet | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 0.994 ms | 34 - 34 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.dlc) |
68
+ | GoogLeNet | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.056 ms | 13 - 13 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet.onnx.zip) |
69
+ | GoogLeNet | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 0.867 ms | 0 - 22 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
70
+ | GoogLeNet | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 0.802 ms | 0 - 23 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
71
+ | GoogLeNet | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 0.338 ms | 0 - 36 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
72
+ | GoogLeNet | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 0.437 ms | 0 - 38 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
73
+ | GoogLeNet | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 0.271 ms | 0 - 40 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
74
+ | GoogLeNet | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 0.247 ms | 0 - 40 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
75
+ | GoogLeNet | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 0.532 ms | 0 - 50 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
76
+ | GoogLeNet | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 0.475 ms | 0 - 22 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
77
+ | GoogLeNet | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 0.435 ms | 0 - 23 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
78
+ | GoogLeNet | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 0.95 ms | 0 - 31 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
79
+ | GoogLeNet | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 1.097 ms | 0 - 31 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
80
+ | GoogLeNet | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 9.638 ms | 9 - 25 MB | CPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
81
+ | GoogLeNet | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 6.196 ms | 0 - 2 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
82
+ | GoogLeNet | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 8.37 ms | 9 - 18 MB | CPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
83
+ | GoogLeNet | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 0.867 ms | 0 - 22 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
84
+ | GoogLeNet | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 0.802 ms | 0 - 23 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
85
+ | GoogLeNet | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 0.273 ms | 0 - 40 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
86
+ | GoogLeNet | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 0.237 ms | 0 - 40 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
87
+ | GoogLeNet | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 0.648 ms | 0 - 28 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
88
+ | GoogLeNet | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 0.59 ms | 0 - 29 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
89
+ | GoogLeNet | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 0.264 ms | 0 - 40 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
90
+ | GoogLeNet | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 0.251 ms | 0 - 40 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
91
+ | GoogLeNet | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 0.475 ms | 0 - 22 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
92
+ | GoogLeNet | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 0.435 ms | 0 - 23 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
93
+ | GoogLeNet | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.197 ms | 0 - 39 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
94
+ | GoogLeNet | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.188 ms | 0 - 36 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
95
+ | GoogLeNet | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 0.366 ms | 0 - 41 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
96
+ | GoogLeNet | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 0.163 ms | 0 - 30 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
97
+ | GoogLeNet | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.153 ms | 0 - 28 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
98
+ | GoogLeNet | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 0.319 ms | 0 - 29 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
99
+ | GoogLeNet | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | TFLITE | 0.351 ms | 0 - 30 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
100
+ | GoogLeNet | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 0.334 ms | 0 - 35 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
101
+ | GoogLeNet | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | ONNX | 9.627 ms | 10 - 26 MB | CPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
102
+ | GoogLeNet | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 0.141 ms | 0 - 24 MB | NPU | [GoogLeNet.tflite](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.tflite) |
103
+ | GoogLeNet | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 0.139 ms | 0 - 25 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
104
+ | GoogLeNet | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 0.314 ms | 0 - 34 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
105
+ | GoogLeNet | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 0.347 ms | 29 - 29 MB | NPU | [GoogLeNet.dlc](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.dlc) |
106
+ | GoogLeNet | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 0.458 ms | 8 - 8 MB | NPU | [GoogLeNet.onnx.zip](https://huggingface.co/qualcomm/GoogLeNet/blob/main/GoogLeNet_w8a8.onnx.zip) |
107
 
108
 
109
 
 
117
  ```
118
 
119
 
120
+ ## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
121
 
122
+ Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.com/) with your
123
  Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
124
 
125
  With this API token, you can configure your client to run models on the cloud
 
127
  ```bash
128
  qai-hub configure --api_token API_TOKEN
129
  ```
130
+ Navigate to [docs](https://workbench.aihub.qualcomm.com/docs/) for more information.
131
 
132
 
133
 
 
238
  spot check the output with expected output.
239
 
240
  **Note**: This on-device profiling and inference requires access to Qualcomm®
241
+ AI Hub Workbench. [Sign up for access](https://myaccount.qualcomm.com/signup).
242
 
243
 
244