qaihm-bot commited on
Commit
c5eeb92
·
verified ·
1 Parent(s): 698e533

See https://github.com/quic/ai-hub-models/releases/v0.34.0 for changelog.

README.md CHANGED
@@ -24,6 +24,7 @@ More details on model performance across various devices, can be found
24
  [here](https://aihub.qualcomm.com/models/wideresnet50).
25
 
26
 
 
27
  ### Model Details
28
 
29
  - **Model Type:** Model_use_case.image_classification
@@ -36,66 +37,66 @@ 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
- | WideResNet50 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 24.01 ms | 0 - 92 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
40
  | WideResNet50 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 24.287 ms | 1 - 41 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
41
- | WideResNet50 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 6.867 ms | 0 - 168 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
42
  | WideResNet50 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 9.726 ms | 1 - 42 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
43
- | WideResNet50 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 4.818 ms | 0 - 1057 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
44
  | WideResNet50 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 4.778 ms | 1 - 12 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
45
- | WideResNet50 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 7.257 ms | 0 - 98 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
46
  | WideResNet50 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 7.082 ms | 1 - 46 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
47
- | WideResNet50 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 24.01 ms | 0 - 92 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
48
  | WideResNet50 | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 24.287 ms | 1 - 41 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
49
- | WideResNet50 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 4.837 ms | 0 - 1022 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
50
  | WideResNet50 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 4.8 ms | 1 - 10 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
51
- | WideResNet50 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 7.867 ms | 0 - 86 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
52
  | WideResNet50 | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 7.647 ms | 0 - 32 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
53
- | WideResNet50 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 4.864 ms | 0 - 955 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
54
  | WideResNet50 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 4.832 ms | 1 - 11 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
55
- | WideResNet50 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 7.257 ms | 0 - 98 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
56
  | WideResNet50 | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 7.082 ms | 1 - 46 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
57
- | WideResNet50 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 4.831 ms | 0 - 1014 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
58
  | WideResNet50 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 4.802 ms | 1 - 11 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
59
  | WideResNet50 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 4.814 ms | 0 - 9 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
60
- | WideResNet50 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 3.557 ms | 0 - 171 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
61
  | WideResNet50 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 3.681 ms | 1 - 53 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
62
  | WideResNet50 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 3.62 ms | 0 - 46 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
63
- | WideResNet50 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 3.397 ms | 0 - 102 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
64
  | WideResNet50 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 3.017 ms | 1 - 45 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
65
  | WideResNet50 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 3.207 ms | 0 - 45 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
66
  | WideResNet50 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 5.1 ms | 465 - 465 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
67
  | WideResNet50 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 4.529 ms | 132 - 132 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
68
- | WideResNet50 | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 3.827 ms | 0 - 37 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
69
- | WideResNet50 | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 3.983 ms | 0 - 38 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
70
- | WideResNet50 | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 2.115 ms | 0 - 106 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
71
- | WideResNet50 | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 2.532 ms | 0 - 105 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
72
- | WideResNet50 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.774 ms | 0 - 387 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
73
- | WideResNet50 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.853 ms | 0 - 378 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
74
- | WideResNet50 | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.924 ms | 0 - 37 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
75
- | WideResNet50 | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.988 ms | 0 - 39 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
76
- | WideResNet50 | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 7.283 ms | 0 - 94 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
77
- | WideResNet50 | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 10.059 ms | 0 - 94 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
78
- | WideResNet50 | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 24.299 ms | 0 - 2 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
79
- | WideResNet50 | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 3.827 ms | 0 - 37 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
80
- | WideResNet50 | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 3.983 ms | 0 - 38 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
81
- | WideResNet50 | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 1.78 ms | 0 - 385 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
82
- | WideResNet50 | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.865 ms | 0 - 380 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
83
- | WideResNet50 | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 2.595 ms | 0 - 39 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
84
- | WideResNet50 | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 2.769 ms | 0 - 43 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
85
- | WideResNet50 | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 1.783 ms | 0 - 389 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
86
- | WideResNet50 | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.865 ms | 0 - 373 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
87
- | WideResNet50 | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 1.924 ms | 0 - 37 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
88
- | WideResNet50 | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.988 ms | 0 - 39 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
89
- | WideResNet50 | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 1.777 ms | 0 - 383 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
90
- | WideResNet50 | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 1.877 ms | 0 - 352 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
91
  | WideResNet50 | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 18.1 ms | 0 - 220 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.onnx) |
92
- | WideResNet50 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.355 ms | 0 - 105 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
93
- | WideResNet50 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.419 ms | 0 - 104 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
94
  | WideResNet50 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 13.625 ms | 6 - 504 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.onnx) |
95
- | WideResNet50 | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 1.218 ms | 0 - 39 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
96
- | WideResNet50 | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 1.221 ms | 0 - 43 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
97
  | WideResNet50 | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 12.123 ms | 5 - 616 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.onnx) |
98
- | WideResNet50 | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.85 ms | 355 - 355 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
99
  | WideResNet50 | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 17.923 ms | 75 - 75 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.onnx) |
100
 
101
 
@@ -154,17 +155,7 @@ device. This script does the following:
154
  ```bash
155
  python -m qai_hub_models.models.wideresnet50.export
156
  ```
157
- ```
158
- Profiling Results
159
- ------------------------------------------------------------
160
- WideResNet50
161
- Device : cs_8275 (ANDROID 14)
162
- Runtime : TFLITE
163
- Estimated inference time (ms) : 24.0
164
- Estimated peak memory usage (MB): [0, 92]
165
- Total # Ops : 79
166
- Compute Unit(s) : npu (79 ops) gpu (0 ops) cpu (0 ops)
167
- ```
168
 
169
 
170
  ## How does this work?
 
24
  [here](https://aihub.qualcomm.com/models/wideresnet50).
25
 
26
 
27
+
28
  ### Model Details
29
 
30
  - **Model Type:** Model_use_case.image_classification
 
37
 
38
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
39
  |---|---|---|---|---|---|---|---|---|
40
+ | WideResNet50 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 23.997 ms | 0 - 98 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
41
  | WideResNet50 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 24.287 ms | 1 - 41 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
42
+ | WideResNet50 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 6.831 ms | 0 - 172 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
43
  | WideResNet50 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 9.726 ms | 1 - 42 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
44
+ | WideResNet50 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 4.836 ms | 0 - 857 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
45
  | WideResNet50 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 4.778 ms | 1 - 12 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
46
+ | WideResNet50 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 7.244 ms | 0 - 98 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
47
  | WideResNet50 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 7.082 ms | 1 - 46 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
48
+ | WideResNet50 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 23.997 ms | 0 - 98 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
49
  | WideResNet50 | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 24.287 ms | 1 - 41 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
50
+ | WideResNet50 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 4.827 ms | 0 - 1034 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
51
  | WideResNet50 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 4.8 ms | 1 - 10 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
52
+ | WideResNet50 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 7.887 ms | 0 - 86 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
53
  | WideResNet50 | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 7.647 ms | 0 - 32 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
54
+ | WideResNet50 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 4.833 ms | 0 - 1037 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
55
  | WideResNet50 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 4.832 ms | 1 - 11 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
56
+ | WideResNet50 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 7.244 ms | 0 - 98 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
57
  | WideResNet50 | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 7.082 ms | 1 - 46 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
58
+ | WideResNet50 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 4.837 ms | 0 - 1005 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
59
  | WideResNet50 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 4.802 ms | 1 - 11 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
60
  | WideResNet50 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 4.814 ms | 0 - 9 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
61
+ | WideResNet50 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 3.591 ms | 0 - 171 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
62
  | WideResNet50 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 3.681 ms | 1 - 53 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
63
  | WideResNet50 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 3.62 ms | 0 - 46 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
64
+ | WideResNet50 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 3.016 ms | 0 - 103 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
65
  | WideResNet50 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 3.017 ms | 1 - 45 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
66
  | WideResNet50 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 3.207 ms | 0 - 45 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
67
  | WideResNet50 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 5.1 ms | 465 - 465 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.dlc) |
68
  | WideResNet50 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 4.529 ms | 132 - 132 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
69
+ | WideResNet50 | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 3.752 ms | 0 - 38 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
70
+ | WideResNet50 | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 3.965 ms | 0 - 39 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
71
+ | WideResNet50 | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 2.164 ms | 0 - 107 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
72
+ | WideResNet50 | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 2.527 ms | 0 - 107 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
73
+ | WideResNet50 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.774 ms | 0 - 389 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
74
+ | WideResNet50 | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.87 ms | 0 - 376 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
75
+ | WideResNet50 | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.911 ms | 0 - 37 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
76
+ | WideResNet50 | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.986 ms | 0 - 39 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
77
+ | WideResNet50 | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 7.575 ms | 0 - 95 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
78
+ | WideResNet50 | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 10.096 ms | 0 - 94 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
79
+ | WideResNet50 | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 24.348 ms | 0 - 2 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
80
+ | WideResNet50 | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 3.752 ms | 0 - 38 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
81
+ | WideResNet50 | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 3.965 ms | 0 - 39 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
82
+ | WideResNet50 | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 1.78 ms | 0 - 389 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
83
+ | WideResNet50 | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.856 ms | 0 - 370 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
84
+ | WideResNet50 | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 2.619 ms | 0 - 40 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
85
+ | WideResNet50 | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 2.751 ms | 0 - 43 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
86
+ | WideResNet50 | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 1.779 ms | 0 - 388 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
87
+ | WideResNet50 | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.871 ms | 0 - 385 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
88
+ | WideResNet50 | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 1.911 ms | 0 - 37 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
89
+ | WideResNet50 | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.986 ms | 0 - 39 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
90
+ | WideResNet50 | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 1.781 ms | 0 - 389 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
91
+ | WideResNet50 | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 1.866 ms | 0 - 368 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
92
  | WideResNet50 | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 18.1 ms | 0 - 220 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.onnx) |
93
+ | WideResNet50 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.369 ms | 0 - 104 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
94
+ | WideResNet50 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.409 ms | 0 - 102 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
95
  | WideResNet50 | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 13.625 ms | 6 - 504 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.onnx) |
96
+ | WideResNet50 | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 1.246 ms | 0 - 39 MB | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.tflite) |
97
+ | WideResNet50 | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 1.264 ms | 0 - 46 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
98
  | WideResNet50 | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 12.123 ms | 5 - 616 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.onnx) |
99
+ | WideResNet50 | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.841 ms | 355 - 355 MB | NPU | [WideResNet50.dlc](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.dlc) |
100
  | WideResNet50 | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 17.923 ms | 75 - 75 MB | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50_w8a8.onnx) |
101
 
102
 
 
155
  ```bash
156
  python -m qai_hub_models.models.wideresnet50.export
157
  ```
158
+
 
 
 
 
 
 
 
 
 
 
159
 
160
 
161
  ## How does this work?
WideResNet50_w8a8.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:69247c012bbaec79ef17b66bee11894963f702e9d1e9f044dd29b9d67be7a08a
3
  size 70979209
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b40968d4c5325d3acbece50d0d3e9a8a7bdf110e72e053e3afb30daa468d8b72
3
  size 70979209
precompiled/qualcomm-qcs6490-proxy/sdk_versions.yml ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ sdk_versions:
2
+ qnn_context_binary:
3
+ qairt: 2.34.2.250528164111_119506
precompiled/qualcomm-snapdragon-x-elite/WideResNet50.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c92e45f76f6411f2509cf2be8b721dcfa6c9e41c20eccc5fc368a202d0db2b16
3
  size 127067543
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7ee521b9dec103291558ada039bf4a7f8c779af14d3041a1ea571386e9db41e
3
  size 127067543
precompiled/qualcomm-snapdragon-x-elite/WideResNet50_w8a8.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:dd9f2c5a67f8452ba342e388ec1e9cf5090edac17bc6ceb691c677e77d3b6f96
3
- size 57323677
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c62c2988c0689c54f045423455edd0fab6be17a8343244e736ae006c7fd7f35
3
+ size 57323678
precompiled/qualcomm-snapdragon-x-elite/sdk_versions.yml ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ sdk_versions:
2
+ qnn_context_binary:
3
+ qairt: 2.34.2.250528164111_119506
4
+ precompiled_qnn_onnx:
5
+ qairt: 2.33.2.250410134701_117956