v0.49.1
Browse filesSee https://github.com/qualcomm/ai-hub-models/releases/v0.49.1 for changelog.
README.md
CHANGED
|
@@ -28,15 +28,15 @@ Below are pre-exported model assets ready for deployment.
|
|
| 28 |
|
| 29 |
| Runtime | Precision | Chipset | SDK Versions | Download |
|
| 30 |
|---|---|---|---|---|
|
| 31 |
-
| ONNX | float | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 32 |
-
| ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 33 |
-
| ONNX | w8a8 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 34 |
-
| ONNX | w8a8_mixed_int16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 35 |
-
| QNN_DLC | float | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 36 |
-
| QNN_DLC | w8a16 | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 37 |
-
| QNN_DLC | w8a8 | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 38 |
-
| TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 39 |
-
| TFLITE | w8a8 | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.
|
| 40 |
|
| 41 |
For more device-specific assets and performance metrics, visit **[VIT on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/vit)**.
|
| 42 |
|
|
@@ -67,78 +67,78 @@ See our repository for [VIT on GitHub](https://github.com/qualcomm/ai-hub-models
|
|
| 67 |
## Performance Summary
|
| 68 |
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|
| 69 |
|---|---|---|---|---|---|---
|
| 70 |
-
| VIT | ONNX | float | Snapdragon®
|
| 71 |
-
| VIT | ONNX | float | Snapdragon®
|
| 72 |
-
| VIT | ONNX | float | Snapdragon®
|
| 73 |
-
| VIT | ONNX | float |
|
| 74 |
-
| VIT | ONNX | float | Qualcomm®
|
| 75 |
-
| VIT | ONNX | float |
|
| 76 |
-
| VIT | ONNX | float | Snapdragon® 8 Elite
|
| 77 |
-
| VIT | ONNX | w8a16 | Snapdragon®
|
| 78 |
-
| VIT | ONNX | w8a16 | Snapdragon®
|
| 79 |
-
| VIT | ONNX | w8a16 | Snapdragon®
|
| 80 |
-
| VIT | ONNX | w8a16 |
|
| 81 |
-
| VIT | ONNX | w8a16 | Qualcomm®
|
| 82 |
-
| VIT | ONNX | w8a16 | Qualcomm®
|
| 83 |
-
| VIT | ONNX | w8a16 | Qualcomm®
|
| 84 |
-
| VIT | ONNX | w8a16 |
|
| 85 |
-
| VIT | ONNX | w8a16 | Snapdragon®
|
| 86 |
-
| VIT | ONNX | w8a16 | Snapdragon®
|
| 87 |
-
| VIT | ONNX | w8a8 | Snapdragon®
|
| 88 |
-
| VIT | ONNX | w8a8 | Snapdragon®
|
| 89 |
-
| VIT | ONNX | w8a8 | Snapdragon®
|
| 90 |
-
| VIT | ONNX | w8a8 |
|
| 91 |
-
| VIT | ONNX | w8a8 | Qualcomm®
|
| 92 |
-
| VIT | ONNX | w8a8 | Qualcomm®
|
| 93 |
-
| VIT | ONNX | w8a8 | Qualcomm®
|
| 94 |
-
| VIT | ONNX | w8a8 |
|
| 95 |
-
| VIT | ONNX | w8a8 | Snapdragon®
|
| 96 |
-
| VIT | ONNX | w8a8 | Snapdragon®
|
| 97 |
-
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon®
|
| 98 |
-
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon®
|
| 99 |
-
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon®
|
| 100 |
-
| VIT | ONNX | w8a8_mixed_int16 |
|
| 101 |
-
| VIT | ONNX | w8a8_mixed_int16 | Qualcomm®
|
| 102 |
-
| VIT | ONNX | w8a8_mixed_int16 | Qualcomm®
|
| 103 |
-
| VIT | ONNX | w8a8_mixed_int16 | Qualcomm®
|
| 104 |
-
| VIT | ONNX | w8a8_mixed_int16 |
|
| 105 |
-
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon®
|
| 106 |
-
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon®
|
| 107 |
-
| VIT | QNN_DLC | float | Snapdragon®
|
| 108 |
-
| VIT | QNN_DLC | float | Snapdragon®
|
| 109 |
-
| VIT | QNN_DLC | float | Snapdragon®
|
| 110 |
-
| VIT | QNN_DLC | float |
|
| 111 |
-
| VIT | QNN_DLC | float | Qualcomm®
|
| 112 |
-
| VIT | QNN_DLC | float | Qualcomm®
|
| 113 |
-
| VIT | QNN_DLC | float | Qualcomm®
|
| 114 |
-
| VIT | QNN_DLC | float | Qualcomm®
|
| 115 |
-
| VIT | QNN_DLC | float | Qualcomm®
|
| 116 |
-
| VIT | QNN_DLC | float | Qualcomm®
|
| 117 |
-
| VIT | QNN_DLC | float |
|
| 118 |
-
| VIT | QNN_DLC | float | Snapdragon® 8 Elite
|
| 119 |
-
| VIT | TFLITE | float | Snapdragon® 8 Gen
|
| 120 |
-
| VIT | TFLITE | float |
|
| 121 |
-
| VIT | TFLITE | float | Qualcomm®
|
| 122 |
-
| VIT | TFLITE | float | Qualcomm®
|
| 123 |
-
| VIT | TFLITE | float | Qualcomm®
|
| 124 |
-
| VIT | TFLITE | float | Qualcomm®
|
| 125 |
-
| VIT | TFLITE | float | Qualcomm®
|
| 126 |
-
| VIT | TFLITE | float | Qualcomm®
|
| 127 |
-
| VIT | TFLITE | float |
|
| 128 |
-
| VIT | TFLITE | float | Snapdragon® 8 Elite
|
| 129 |
-
| VIT | TFLITE | w8a8 | Snapdragon® 8 Gen
|
| 130 |
-
| VIT | TFLITE | w8a8 |
|
| 131 |
-
| VIT | TFLITE | w8a8 | Qualcomm®
|
| 132 |
-
| VIT | TFLITE | w8a8 | Qualcomm®
|
| 133 |
-
| VIT | TFLITE | w8a8 | Qualcomm®
|
|
|
|
| 134 |
| VIT | TFLITE | w8a8 | Qualcomm® QCS9075 | 7.561 ms | 0 - 89 MB | NPU
|
| 135 |
-
| VIT | TFLITE | w8a8 | Qualcomm® QCM6690 |
|
| 136 |
-
| VIT | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 8.
|
| 137 |
-
| VIT | TFLITE | w8a8 | Qualcomm® SA7255P | 14.
|
| 138 |
-
| VIT | TFLITE | w8a8 | Qualcomm® SA8295P | 9.
|
| 139 |
-
| VIT | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 3.
|
| 140 |
-
| VIT | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 20.
|
| 141 |
-
| VIT | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 2.29 ms | 0 - 90 MB | NPU
|
| 142 |
|
| 143 |
## License
|
| 144 |
* The license for the original implementation of VIT can be found
|
|
|
|
| 28 |
|
| 29 |
| Runtime | Precision | Chipset | SDK Versions | Download |
|
| 30 |
|---|---|---|---|---|
|
| 31 |
+
| ONNX | float | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-onnx-float.zip)
|
| 32 |
+
| ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-onnx-w8a16.zip)
|
| 33 |
+
| ONNX | w8a8 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-onnx-w8a8.zip)
|
| 34 |
+
| ONNX | w8a8_mixed_int16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-onnx-w8a8_mixed_int16.zip)
|
| 35 |
+
| QNN_DLC | float | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-qnn_dlc-float.zip)
|
| 36 |
+
| QNN_DLC | w8a16 | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-qnn_dlc-w8a16.zip)
|
| 37 |
+
| QNN_DLC | w8a8 | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-qnn_dlc-w8a8.zip)
|
| 38 |
+
| TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-tflite-float.zip)
|
| 39 |
+
| TFLITE | w8a8 | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/releases/v0.49.1/vit-tflite-w8a8.zip)
|
| 40 |
|
| 41 |
For more device-specific assets and performance metrics, visit **[VIT on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/vit)**.
|
| 42 |
|
|
|
|
| 67 |
## Performance Summary
|
| 68 |
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|
| 69 |
|---|---|---|---|---|---|---
|
| 70 |
+
| VIT | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.667 ms | 1 - 355 MB | NPU
|
| 71 |
+
| VIT | ONNX | float | Snapdragon® X2 Elite | 3.876 ms | 170 - 170 MB | NPU
|
| 72 |
+
| VIT | ONNX | float | Snapdragon® X Elite | 11.109 ms | 170 - 170 MB | NPU
|
| 73 |
+
| VIT | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 7.167 ms | 0 - 377 MB | NPU
|
| 74 |
+
| VIT | ONNX | float | Qualcomm® QCS8550 (Proxy) | 10.487 ms | 0 - 195 MB | NPU
|
| 75 |
+
| VIT | ONNX | float | Qualcomm® QCS9075 | 14.321 ms | 0 - 4 MB | NPU
|
| 76 |
+
| VIT | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.987 ms | 0 - 344 MB | NPU
|
| 77 |
+
| VIT | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 3.797 ms | 0 - 301 MB | NPU
|
| 78 |
+
| VIT | ONNX | w8a16 | Snapdragon® X2 Elite | 3.894 ms | 86 - 86 MB | NPU
|
| 79 |
+
| VIT | ONNX | w8a16 | Snapdragon® X Elite | 11.261 ms | 86 - 86 MB | NPU
|
| 80 |
+
| VIT | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 7.372 ms | 0 - 379 MB | NPU
|
| 81 |
+
| VIT | ONNX | w8a16 | Qualcomm® QCS6490 | 1108.405 ms | 35 - 60 MB | CPU
|
| 82 |
+
| VIT | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 10.713 ms | 0 - 477 MB | NPU
|
| 83 |
+
| VIT | ONNX | w8a16 | Qualcomm® QCS9075 | 12.785 ms | 0 - 3 MB | NPU
|
| 84 |
+
| VIT | ONNX | w8a16 | Qualcomm® QCM6690 | 614.255 ms | 73 - 88 MB | CPU
|
| 85 |
+
| VIT | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 5.306 ms | 0 - 295 MB | NPU
|
| 86 |
+
| VIT | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 594.848 ms | 86 - 107 MB | CPU
|
| 87 |
+
| VIT | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 4.581 ms | 0 - 351 MB | NPU
|
| 88 |
+
| VIT | ONNX | w8a8 | Snapdragon® X2 Elite | 5.106 ms | 85 - 85 MB | NPU
|
| 89 |
+
| VIT | ONNX | w8a8 | Snapdragon® X Elite | 13.631 ms | 85 - 85 MB | NPU
|
| 90 |
+
| VIT | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 8.8 ms | 0 - 466 MB | NPU
|
| 91 |
+
| VIT | ONNX | w8a8 | Qualcomm® QCS6490 | 354.711 ms | 21 - 79 MB | CPU
|
| 92 |
+
| VIT | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 12.932 ms | 0 - 100 MB | NPU
|
| 93 |
+
| VIT | ONNX | w8a8 | Qualcomm® QCS9075 | 13.588 ms | 0 - 3 MB | NPU
|
| 94 |
+
| VIT | ONNX | w8a8 | Qualcomm® QCM6690 | 135.114 ms | 14 - 32 MB | CPU
|
| 95 |
+
| VIT | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 7.277 ms | 0 - 319 MB | NPU
|
| 96 |
+
| VIT | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 128.787 ms | 24 - 43 MB | CPU
|
| 97 |
+
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Elite Gen 5 Mobile | 63.145 ms | 0 - 268 MB | NPU
|
| 98 |
+
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon® X2 Elite | 53.502 ms | 79 - 79 MB | NPU
|
| 99 |
+
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon® X Elite | 174.304 ms | 79 - 79 MB | NPU
|
| 100 |
+
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Gen 3 Mobile | 81.808 ms | 67 - 429 MB | NPU
|
| 101 |
+
| VIT | ONNX | w8a8_mixed_int16 | Qualcomm® QCS6490 | 728.091 ms | 97 - 126 MB | CPU
|
| 102 |
+
| VIT | ONNX | w8a8_mixed_int16 | Qualcomm® QCS8550 (Proxy) | 101.273 ms | 0 - 382 MB | NPU
|
| 103 |
+
| VIT | ONNX | w8a8_mixed_int16 | Qualcomm® QCS9075 | 130.774 ms | 68 - 70 MB | NPU
|
| 104 |
+
| VIT | ONNX | w8a8_mixed_int16 | Qualcomm® QCM6690 | 387.043 ms | 103 - 124 MB | CPU
|
| 105 |
+
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon® 8 Elite For Galaxy Mobile | 88.161 ms | 68 - 340 MB | NPU
|
| 106 |
+
| VIT | ONNX | w8a8_mixed_int16 | Snapdragon® 7 Gen 4 Mobile | 372.061 ms | 35 - 55 MB | CPU
|
| 107 |
+
| VIT | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 4.052 ms | 1 - 344 MB | NPU
|
| 108 |
+
| VIT | QNN_DLC | float | Snapdragon® X2 Elite | 4.478 ms | 1 - 1 MB | NPU
|
| 109 |
+
| VIT | QNN_DLC | float | Snapdragon® X Elite | 11.859 ms | 1 - 1 MB | NPU
|
| 110 |
+
| VIT | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 7.716 ms | 0 - 369 MB | NPU
|
| 111 |
+
| VIT | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 40.633 ms | 1 - 339 MB | NPU
|
| 112 |
+
| VIT | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 11.125 ms | 1 - 407 MB | NPU
|
| 113 |
+
| VIT | QNN_DLC | float | Qualcomm® SA8775P | 64.028 ms | 1 - 339 MB | NPU
|
| 114 |
+
| VIT | QNN_DLC | float | Qualcomm® QCS9075 | 14.893 ms | 1 - 3 MB | NPU
|
| 115 |
+
| VIT | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 19.096 ms | 0 - 350 MB | NPU
|
| 116 |
+
| VIT | QNN_DLC | float | Qualcomm® SA7255P | 40.633 ms | 1 - 339 MB | NPU
|
| 117 |
+
| VIT | QNN_DLC | float | Qualcomm® SA8295P | 17.182 ms | 1 - 332 MB | NPU
|
| 118 |
+
| VIT | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 5.295 ms | 0 - 338 MB | NPU
|
| 119 |
+
| VIT | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.094 ms | 0 - 278 MB | NPU
|
| 120 |
+
| VIT | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 5.87 ms | 0 - 318 MB | NPU
|
| 121 |
+
| VIT | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 35.876 ms | 0 - 288 MB | NPU
|
| 122 |
+
| VIT | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 8.002 ms | 0 - 3 MB | NPU
|
| 123 |
+
| VIT | TFLITE | float | Qualcomm® SA8775P | 11.156 ms | 0 - 288 MB | NPU
|
| 124 |
+
| VIT | TFLITE | float | Qualcomm® QCS9075 | 11.649 ms | 0 - 174 MB | NPU
|
| 125 |
+
| VIT | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 13.857 ms | 0 - 292 MB | NPU
|
| 126 |
+
| VIT | TFLITE | float | Qualcomm® SA7255P | 35.876 ms | 0 - 288 MB | NPU
|
| 127 |
+
| VIT | TFLITE | float | Qualcomm® SA8295P | 13.383 ms | 0 - 262 MB | NPU
|
| 128 |
+
| VIT | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.943 ms | 0 - 291 MB | NPU
|
| 129 |
+
| VIT | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 2.295 ms | 0 - 88 MB | NPU
|
| 130 |
+
| VIT | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 4.735 ms | 0 - 182 MB | NPU
|
| 131 |
+
| VIT | TFLITE | w8a8 | Qualcomm® QCS6490 | 78.182 ms | 1 - 99 MB | NPU
|
| 132 |
+
| VIT | TFLITE | w8a8 | Qualcomm® QCS8275 (Proxy) | 14.341 ms | 0 - 85 MB | NPU
|
| 133 |
+
| VIT | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 6.776 ms | 0 - 4 MB | NPU
|
| 134 |
+
| VIT | TFLITE | w8a8 | Qualcomm® SA8775P | 7.074 ms | 0 - 86 MB | NPU
|
| 135 |
| VIT | TFLITE | w8a8 | Qualcomm® QCS9075 | 7.561 ms | 0 - 89 MB | NPU
|
| 136 |
+
| VIT | TFLITE | w8a8 | Qualcomm® QCM6690 | 101.681 ms | 1 - 185 MB | NPU
|
| 137 |
+
| VIT | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 8.676 ms | 0 - 181 MB | NPU
|
| 138 |
+
| VIT | TFLITE | w8a8 | Qualcomm® SA7255P | 14.341 ms | 0 - 85 MB | NPU
|
| 139 |
+
| VIT | TFLITE | w8a8 | Qualcomm® SA8295P | 9.695 ms | 0 - 89 MB | NPU
|
| 140 |
+
| VIT | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 3.374 ms | 0 - 84 MB | NPU
|
| 141 |
+
| VIT | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 20.157 ms | 1 - 74 MB | NPU
|
|
|
|
| 142 |
|
| 143 |
## License
|
| 144 |
* The license for the original implementation of VIT can be found
|