qaihm-bot commited on
Commit
1d6118b
·
verified ·
1 Parent(s): 5caae20

See https://github.com/quic/ai-hub-models/releases/v0.38.0 for changelog.

README.md CHANGED
@@ -37,57 +37,52 @@ More details on model performance across various devices, can be found
37
 
38
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
39
  |---|---|---|---|---|---|---|---|---|
40
- | Swin-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 60.88 ms | 0 - 356 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
41
- | Swin-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 54.394 ms | 0 - 316 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
42
- | Swin-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 28.423 ms | 0 - 351 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
43
- | Swin-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 28.209 ms | 1 - 330 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
44
- | Swin-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 21.901 ms | 0 - 31 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
45
- | Swin-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 19.28 ms | 0 - 65 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
46
- | Swin-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 19.37 ms | 0 - 62 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
47
- | Swin-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 25.164 ms | 0 - 355 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
48
- | Swin-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 22.412 ms | 0 - 314 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
49
- | Swin-Base | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 60.88 ms | 0 - 356 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
50
- | Swin-Base | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 54.394 ms | 0 - 316 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
51
- | Swin-Base | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 21.708 ms | 0 - 32 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
52
- | Swin-Base | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 19.559 ms | 0 - 70 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
53
- | Swin-Base | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 31.692 ms | 0 - 347 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
54
- | Swin-Base | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 28.368 ms | 1 - 306 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
55
- | Swin-Base | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 21.818 ms | 0 - 36 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
56
- | Swin-Base | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 19.339 ms | 0 - 63 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
57
- | Swin-Base | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 25.164 ms | 0 - 355 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
58
- | Swin-Base | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 22.412 ms | 0 - 314 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
59
- | Swin-Base | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 22.016 ms | 0 - 36 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
60
- | Swin-Base | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 19.473 ms | 0 - 61 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
61
- | Swin-Base | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 19.321 ms | 0 - 252 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
62
- | Swin-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 15.329 ms | 0 - 357 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
63
- | Swin-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 13.32 ms | 1 - 343 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
64
- | Swin-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 13.24 ms | 1 - 340 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
65
- | Swin-Base | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 14.075 ms | 0 - 348 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
66
- | Swin-Base | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 9.776 ms | 1 - 312 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
67
- | Swin-Base | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 11.828 ms | 1 - 309 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
68
- | Swin-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 20.028 ms | 1025 - 1025 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
69
- | Swin-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 19.655 ms | 175 - 175 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
70
- | Swin-Base | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 34.662 ms | 14 - 259 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
71
- | Swin-Base | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 23.526 ms | 0 - 249 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
72
- | Swin-Base | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 17.937 ms | 0 - 67 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
73
- | Swin-Base | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 178.294 ms | 484 - 756 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
74
- | Swin-Base | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 18.318 ms | 0 - 243 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
75
- | Swin-Base | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 76.576 ms | 0 - 946 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
76
- | Swin-Base | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 553.953 ms | 109 - 135 MB | CPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
77
- | Swin-Base | w8a16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 687.751 ms | 70 - 154 MB | CPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
78
- | Swin-Base | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 34.662 ms | 14 - 259 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
79
- | Swin-Base | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 18.044 ms | 0 - 54 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
80
- | Swin-Base | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 21.976 ms | 0 - 249 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
81
- | Swin-Base | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 18.048 ms | 0 - 55 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
82
- | Swin-Base | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 18.318 ms | 0 - 243 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
83
- | Swin-Base | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 18.071 ms | 0 - 71 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
84
- | Swin-Base | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 191.242 ms | 463 - 715 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
85
- | Swin-Base | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 12.112 ms | 0 - 249 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
86
- | Swin-Base | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 140.637 ms | 640 - 974 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
87
- | Swin-Base | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 9.437 ms | 0 - 242 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
88
- | Swin-Base | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 109.395 ms | 615 - 888 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
89
- | Swin-Base | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 18.968 ms | 291 - 291 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
90
- | Swin-Base | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 150.463 ms | 989 - 989 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
91
 
92
 
93
 
@@ -169,7 +164,7 @@ from qai_hub_models.models.swin_base import Model
169
  torch_model = Model.from_pretrained()
170
 
171
  # Device
172
- device = hub.Device("Samsung Galaxy S24")
173
 
174
  # Trace model
175
  input_shape = torch_model.get_input_spec()
 
37
 
38
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
39
  |---|---|---|---|---|---|---|---|---|
40
+ | Swin-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 60.981 ms | 0 - 356 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
41
+ | Swin-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 55.327 ms | 0 - 314 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
42
+ | Swin-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 28.617 ms | 0 - 353 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
43
+ | Swin-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 29.331 ms | 0 - 337 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
44
+ | Swin-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 21.543 ms | 0 - 37 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
45
+ | Swin-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 19.841 ms | 0 - 48 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
46
+ | Swin-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 19.241 ms | 0 - 32 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
47
+ | Swin-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 25.098 ms | 0 - 355 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
48
+ | Swin-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 22.891 ms | 0 - 315 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
49
+ | Swin-Base | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 60.981 ms | 0 - 356 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
50
+ | Swin-Base | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 55.327 ms | 0 - 314 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
51
+ | Swin-Base | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 21.756 ms | 0 - 38 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
52
+ | Swin-Base | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 19.716 ms | 0 - 67 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
53
+ | Swin-Base | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 31.811 ms | 0 - 346 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
54
+ | Swin-Base | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 29.88 ms | 0 - 308 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
55
+ | Swin-Base | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 21.8 ms | 0 - 32 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
56
+ | Swin-Base | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 19.818 ms | 0 - 48 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
57
+ | Swin-Base | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 25.098 ms | 0 - 355 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
58
+ | Swin-Base | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 22.891 ms | 0 - 315 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
59
+ | Swin-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 15.335 ms | 0 - 358 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
60
+ | Swin-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 13.312 ms | 1 - 347 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
61
+ | Swin-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 12.826 ms | 1 - 347 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
62
+ | Swin-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 12.005 ms | 0 - 354 MB | NPU | [Swin-Base.tflite](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.tflite) |
63
+ | Swin-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 10.213 ms | 1 - 313 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
64
+ | Swin-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 9.64 ms | 1 - 316 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
65
+ | Swin-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 20.368 ms | 1031 - 1031 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.dlc) |
66
+ | Swin-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 19.536 ms | 175 - 175 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base.onnx.zip) |
67
+ | Swin-Base | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 35.765 ms | 0 - 283 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
68
+ | Swin-Base | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 25.56 ms | 0 - 291 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
69
+ | Swin-Base | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 18.784 ms | 0 - 67 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
70
+ | Swin-Base | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 151.45 ms | 524 - 818 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
71
+ | Swin-Base | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 18.934 ms | 0 - 285 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
72
+ | Swin-Base | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 73.541 ms | 0 - 1006 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
73
+ | Swin-Base | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 556.075 ms | 112 - 138 MB | CPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
74
+ | Swin-Base | w8a16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 607.408 ms | 70 - 166 MB | CPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
75
+ | Swin-Base | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 35.765 ms | 0 - 283 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
76
+ | Swin-Base | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 18.532 ms | 0 - 87 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
77
+ | Swin-Base | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 22.993 ms | 0 - 287 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
78
+ | Swin-Base | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 18.766 ms | 0 - 78 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
79
+ | Swin-Base | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 18.934 ms | 0 - 285 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
80
+ | Swin-Base | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 12.658 ms | 0 - 291 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
81
+ | Swin-Base | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 126.161 ms | 643 - 1011 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
82
+ | Swin-Base | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 10.227 ms | 0 - 273 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
83
+ | Swin-Base | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 99.929 ms | 663 - 958 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
84
+ | Swin-Base | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 19.437 ms | 354 - 354 MB | NPU | [Swin-Base.dlc](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.dlc) |
85
+ | Swin-Base | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 153.494 ms | 987 - 987 MB | NPU | [Swin-Base.onnx.zip](https://huggingface.co/qualcomm/Swin-Base/blob/main/Swin-Base_w8a16.onnx.zip) |
 
 
 
 
 
86
 
87
 
88
 
 
164
  torch_model = Model.from_pretrained()
165
 
166
  # Device
167
+ device = hub.Device("Samsung Galaxy S25")
168
 
169
  # Trace model
170
  input_shape = torch_model.get_input_spec()
Swin-Base_float.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5515fe4db9cf2d313e3d25f3c2f3937c648f4bec204a0266f710d7fe18a7c310
3
- size 356144300
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04ac17e3ee619fe3405e40b716872aef0ae1ff19b2fb49d9458a343aaa14f63d
3
+ size 356143796
Swin-Base_float.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:18975bb79b3f3eb2444272b775255d85080856ce4b70eff9584bd5913dde6243
3
- size 326769875
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b281add64c46700cebb2f506090441b504c410ea414b9bfbca1b2c9aae57890
3
+ size 326769699
Swin-Base_w8a16.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3f39f8717b55ff102124f22017dc67dd24e1be6880be795eb523b86dda75146c
3
- size 94624508
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99adae14077fa6ea1ac01d0b009dabccd8c86a939f9ed2f74f8069735c3a6af7
3
+ size 94623940
Swin-Base_w8a16.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5e57d6a3e2378db97d794b9afcaafca25328f28e9bd980ed43585fd3b8df9b73
3
- size 307805567
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a869f8c18bd380eafbef7abcfb7ec60655b713c5fb43247c9c800d6ed570d22b
3
+ size 307803669
precompiled/qualcomm-qcs6490-proxy/Swin-Base_w8a16.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:3ee39aa79696379bdb3557c725f4c2284474ca4d1e7b77500d4078ed9daa8e44
3
- size 96071680
 
 
 
 
precompiled/qualcomm-qcs6490-proxy/tool-versions.yaml DELETED
@@ -1,3 +0,0 @@
1
- tool_versions:
2
- qnn_context_binary:
3
- qairt: 2.37.0.250724175447_124859
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/Swin-Base_float.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:bc50a6ab1f1a8751da1ab312b10edf07367fc0310101c3dc682dce6bacf950b9
3
- size 182857728
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/Swin-Base_float.onnx.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:222061132e1d7dc729969ad384f9cc7e3065d881649f9a63d8d307e7664bddb3
3
- size 164692120
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/Swin-Base_w8a16.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:288b28a41a8af2f4c533af421c9b9ff420ed8bcdc4faf9a3703535aa83c86387
3
- size 96186368
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/tool-versions.yaml DELETED
@@ -1,3 +0,0 @@
1
- tool_versions:
2
- qnn_context_binary:
3
- qairt: 2.37.0.250724175447_124859
 
 
 
 
tool-versions.yaml CHANGED
@@ -1,4 +1,4 @@
1
  tool_versions:
2
  onnx:
3
- qairt: 2.36.4.250725200057_123280
4
  onnx_runtime: 1.22.2
 
1
  tool_versions:
2
  onnx:
3
+ qairt: 2.37.1.250807093845_124904
4
  onnx_runtime: 1.22.2