Hugging Face's logo

qualcomm
/
LeViT

Image Classification
PyTorch
LiteRT
bu_auto
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
6dccd19
·
verified ·
1 Parent(s): 7de85f8

v0.38.0

Browse files

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

Files changed (16) hide show
  1. LeViT_float.onnx.zip +2 -2
  2. LeViT_float.tflite +2 -2
  3. LeViT_w8a16.dlc +2 -2
  4. LeViT_w8a16.onnx.zip +2 -2
  5. LeViT_w8a16_mixed_int16.dlc +2 -2
  6. LeViT_w8a16_mixed_int16.onnx.zip +2 -2
  7. README.md +37 -43
  8. precompiled/qualcomm-qcs6490-proxy/LeViT_w8a16.bin +0 -3
  9. precompiled/qualcomm-qcs6490-proxy/LeViT_w8a16.onnx.zip +0 -3
  10. precompiled/qualcomm-qcs6490-proxy/tool-versions.yaml +0 -3
  11. precompiled/qualcomm-snapdragon-x-elite/LeViT_w8a16.bin +0 -3
  12. precompiled/qualcomm-snapdragon-x-elite/LeViT_w8a16.onnx.zip +0 -3
  13. precompiled/qualcomm-snapdragon-x-elite/LeViT_w8a16_mixed_int16.bin +0 -3
  14. precompiled/qualcomm-snapdragon-x-elite/LeViT_w8a16_mixed_int16.onnx.zip +0 -3
  15. precompiled/qualcomm-snapdragon-x-elite/tool-versions.yaml +0 -3
  16. tool-versions.yaml +1 -1
LeViT_float.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9c142c623c8473bf713369d32417f4fbfb70b7ccf1261f3b6b83267b210ca397
3
- size 27621133
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25d55b9f7443b91d393fa42e777ff46ec78687a28eee424c257efbaab3c2001f
3
+ size 27621067
LeViT_float.tflite CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:13770b994b8d8480d84b217fa508e1fa1b4537f1557292ecf4ef0dec9f701a01
3
- size 31342312
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12008a07e39d199527251c22951082704cfd754b1b573b098200cb00a707b3a2
3
+ size 31342316
LeViT_w8a16.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:dc5908dbb8ca123e7f62c28518875a296516c1b96b1466fcccff2074a5a48172
3
- size 8619004
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dba38cfe1f27255f04f72912dce3756c6670ebd41c0c354a58194cff26ddfba9
3
+ size 8620548
LeViT_w8a16.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e081179dbf99d31f362b2a1ffefd994b8047e0e1d47b559ad9b1b3635752af5f
3
- size 10499355
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd9e6320fc9d5c45af3b2bf2c4f2878dcfb41aa6402523ba5a7f882e52953aae
3
+ size 10499779
LeViT_w8a16_mixed_int16.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3ad951629a4618159192533050bf4588c39a6716107e353019d42813d8d596ca
3
- size 9934332
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d3d8a0ac0f49d8d3af82a5c2f29512b38b355a3f04d9d313b28d0e4ee6a67bec
3
+ size 9935876
LeViT_w8a16_mixed_int16.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:39edd91c668367264265ca6a89edd142c5ed3d0d864d79c21f878ed8cd1b32fe
3
- size 12174870
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b81cd83de4d51f0172ddd3ee320d11256d919734e01ed597c69d26ec181029b1
3
+ size 12175302
README.md CHANGED
@@ -36,48 +36,42 @@ 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
- | LeViT | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 4.127 ms | 0 - 42 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
40
- | LeViT | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.814 ms | 0 - 51 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
41
- | LeViT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.542 ms | 0 - 91 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
42
- | LeViT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 1.643 ms | 0 - 66 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
43
- | LeViT | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 2.06 ms | 0 - 42 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
44
- | LeViT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 1.564 ms | 0 - 88 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
45
- | LeViT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.655 ms | 0 - 73 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
46
- | LeViT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.066 ms | 0 - 54 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
47
- | LeViT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.085 ms | 0 - 50 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
48
- | LeViT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 1.017 ms | 0 - 47 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
49
- | LeViT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.147 ms | 1 - 46 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
50
- | LeViT | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.62 ms | 16 - 16 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
51
- | LeViT | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 2.741 ms | 0 - 25 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
52
- | LeViT | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.644 ms | 0 - 35 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
53
- | LeViT | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.366 ms | 0 - 9 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
54
- | LeViT | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 3.169 ms | 0 - 31 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
55
- | LeViT | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.678 ms | 0 - 25 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
56
- | LeViT | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 4.268 ms | 0 - 34 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
57
- | LeViT | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 18.383 ms | 6 - 23 MB | CPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
58
- | LeViT | w8a16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 15.124 ms | 5 - 19 MB | CPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
59
- | LeViT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 1.365 ms | 0 - 10 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
60
- | LeViT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 3.159 ms | 0 - 37 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
61
- | LeViT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.935 ms | 0 - 34 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
62
- | LeViT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.19 ms | 0 - 65 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
63
- | LeViT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 0.782 ms | 0 - 26 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
64
- | LeViT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 2.431 ms | 0 - 55 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
65
- | LeViT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.566 ms | 8 - 8 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
66
- | LeViT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.234 ms | 15 - 15 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
67
- | LeViT | w8a16_mixed_int16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 2.801 ms | 0 - 25 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
68
- | LeViT | w8a16_mixed_int16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.372 ms | 0 - 10 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
69
- | LeViT | w8a16_mixed_int16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 4.266 ms | 0 - 36 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
70
- | LeViT | w8a16_mixed_int16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.693 ms | 0 - 25 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
71
- | LeViT | w8a16_mixed_int16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 18.01 ms | 6 - 22 MB | CPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
72
- | LeViT | w8a16_mixed_int16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 14.825 ms | 2 - 17 MB | CPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
73
- | LeViT | w8a16_mixed_int16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 1.369 ms | 0 - 9 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
74
- | LeViT | w8a16_mixed_int16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 3.907 ms | 0 - 40 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
75
- | LeViT | w8a16_mixed_int16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.951 ms | 0 - 35 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
76
- | LeViT | w8a16_mixed_int16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.734 ms | 0 - 71 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
77
- | LeViT | w8a16_mixed_int16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 0.808 ms | 0 - 30 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
78
- | LeViT | w8a16_mixed_int16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 2.915 ms | 0 - 53 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
79
- | LeViT | w8a16_mixed_int16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.576 ms | 21 - 21 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
80
- | LeViT | w8a16_mixed_int16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.866 ms | 12 - 12 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
81
 
82
 
83
 
@@ -159,7 +153,7 @@ from qai_hub_models.models.levit import Model
159
  torch_model = Model.from_pretrained()
160
 
161
  # Device
162
- device = hub.Device("Samsung Galaxy S24")
163
 
164
  # Trace model
165
  input_shape = torch_model.get_input_spec()
 
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
+ | LeViT | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 4.001 ms | 0 - 42 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
40
+ | LeViT | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.785 ms | 0 - 51 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
41
+ | LeViT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.496 ms | 0 - 84 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
42
+ | LeViT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 1.691 ms | 0 - 67 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
43
+ | LeViT | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.973 ms | 0 - 42 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
44
+ | LeViT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.037 ms | 0 - 54 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
45
+ | LeViT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.089 ms | 0 - 52 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
46
+ | LeViT | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 0.807 ms | 0 - 49 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
47
+ | LeViT | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 0.927 ms | 0 - 46 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
48
+ | LeViT | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.602 ms | 16 - 16 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx.zip) |
49
+ | LeViT | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 2.731 ms | 0 - 25 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
50
+ | LeViT | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.594 ms | 0 - 37 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
51
+ | LeViT | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.363 ms | 0 - 11 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
52
+ | LeViT | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 3.077 ms | 0 - 37 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
53
+ | LeViT | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.671 ms | 0 - 25 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
54
+ | LeViT | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 4.262 ms | 0 - 34 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
55
+ | LeViT | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 18.283 ms | 6 - 22 MB | CPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
56
+ | LeViT | w8a16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 14.743 ms | 5 - 19 MB | CPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
57
+ | LeViT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.929 ms | 0 - 37 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
58
+ | LeViT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.166 ms | 0 - 68 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
59
+ | LeViT | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.639 ms | 0 - 31 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
60
+ | LeViT | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 1.987 ms | 0 - 57 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
61
+ | LeViT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.559 ms | 5 - 5 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
62
+ | LeViT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.172 ms | 15 - 15 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx.zip) |
63
+ | LeViT | w8a16_mixed_int16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 2.79 ms | 0 - 25 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
64
+ | LeViT | w8a16_mixed_int16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.372 ms | 0 - 18 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
65
+ | LeViT | w8a16_mixed_int16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 3.893 ms | 0 - 41 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
66
+ | LeViT | w8a16_mixed_int16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.675 ms | 0 - 25 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
67
+ | LeViT | w8a16_mixed_int16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 18.027 ms | 6 - 21 MB | CPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
68
+ | LeViT | w8a16_mixed_int16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 15.108 ms | 5 - 22 MB | CPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
69
+ | LeViT | w8a16_mixed_int16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.948 ms | 0 - 35 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
70
+ | LeViT | w8a16_mixed_int16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.72 ms | 0 - 68 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
71
+ | LeViT | w8a16_mixed_int16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.654 ms | 0 - 30 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
72
+ | LeViT | w8a16_mixed_int16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 2.572 ms | 0 - 57 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
73
+ | LeViT | w8a16_mixed_int16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.577 ms | 15 - 15 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.dlc) |
74
+ | LeViT | w8a16_mixed_int16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.96 ms | 8 - 8 MB | NPU | [LeViT.onnx.zip](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16_mixed_int16.onnx.zip) |
 
 
 
 
 
 
75
 
76
 
77
 
 
153
  torch_model = Model.from_pretrained()
154
 
155
  # Device
156
+ device = hub.Device("Samsung Galaxy S25")
157
 
158
  # Trace model
159
  input_shape = torch_model.get_input_spec()
precompiled/qualcomm-qcs6490-proxy/LeViT_w8a16.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:de5cc953379f5d5d487ccea452b8112e1681537bfa5565c44fe5e7a3ce7c658c
3
- size 8826880
 
 
 
 
precompiled/qualcomm-qcs6490-proxy/LeViT_w8a16.onnx.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:3974cf47d8e7e51d9b4c0f2fd6ce1b9d90aefa773b15fa9c874f173657458931
3
- size 5551850
 
 
 
 
precompiled/qualcomm-qcs6490-proxy/tool-versions.yaml DELETED
@@ -1,3 +0,0 @@
1
- tool_versions:
2
- precompiled_qnn_onnx:
3
- qairt: 2.36.4.250725200057_123280
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/LeViT_w8a16.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:a906cee382ce8acebed34532e77c7d9b7f3cd78fd6530aff0fa9a36e6f86a8e1
3
- size 9048064
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/LeViT_w8a16.onnx.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:518ee914435583399eeb140c0716b7d05ab05873271443f5482d38cc3b76fe06
3
- size 5563026
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/LeViT_w8a16_mixed_int16.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:6097b01ee73bf233ae89d8fea5b91424290b9a1331ffe7b6f438a3a4aa799d23
3
- size 10362880
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/LeViT_w8a16_mixed_int16.onnx.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:1eae96b12c0b39c90e3079efa5d3f8f8adff7c769f558170293fa51395f6f0b2
3
- size 7142264
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/tool-versions.yaml DELETED
@@ -1,3 +0,0 @@
1
- tool_versions:
2
- precompiled_qnn_onnx:
3
- qairt: 2.36.4.250725200057_123280
 
 
 
 
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