Image Classification
PyTorch
bu_auto
android
qaihm-bot commited on
Commit
25cb8a6
·
verified ·
1 Parent(s): 0e29ebd

See https://github.com/quic/ai-hub-models/releases/v0.31.0 for changelog.

Files changed (3) hide show
  1. .gitattributes +1 -0
  2. LeViT_w8a16.dlc +3 -0
  3. README.md +25 -17
.gitattributes CHANGED
@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
  LeViT.so filter=lfs diff=lfs merge=lfs -text
 
 
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
  LeViT.so filter=lfs diff=lfs merge=lfs -text
37
+ LeViT_w8a16.dlc filter=lfs diff=lfs merge=lfs -text
LeViT_w8a16.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1e235397ff4cda15e7e7451a23c61724241104bd42a16b30a61ec7ce7d744a2
3
+ size 9257978
README.md CHANGED
@@ -34,21 +34,29 @@ More details on model performance across various devices, can be found
34
 
35
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
- | LeViT | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 13.633 ms | 0 - 38 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
38
- | LeViT | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.336 ms | 0 - 49 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
39
- | LeViT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.093 ms | 0 - 91 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
40
- | LeViT | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.56 ms | 0 - 42 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
41
- | LeViT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 1.092 ms | 0 - 93 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
42
- | LeViT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.637 ms | 0 - 48 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx) |
43
- | LeViT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.771 ms | 0 - 56 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
44
- | LeViT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.096 ms | 0 - 54 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx) |
45
- | LeViT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 0.761 ms | 0 - 49 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
46
- | LeViT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.0 ms | 1 - 44 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx) |
47
- | LeViT | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.725 ms | 16 - 16 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx) |
48
- | LeViT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 16.899 ms | 2 - 19 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx) |
49
- | LeViT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 14.304 ms | 2 - 32 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx) |
50
- | LeViT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 12.893 ms | 3 - 22 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx) |
51
- | LeViT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 15.395 ms | 8 - 8 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx) |
 
 
 
 
 
 
 
 
52
 
53
 
54
 
@@ -112,8 +120,8 @@ Profiling Results
112
  LeViT
113
  Device : cs_8275 (ANDROID 14)
114
  Runtime : TFLITE
115
- Estimated inference time (ms) : 13.6
116
- Estimated peak memory usage (MB): [0, 38]
117
  Total # Ops : 306
118
  Compute Unit(s) : npu (306 ops) gpu (0 ops) cpu (0 ops)
119
  ```
 
34
 
35
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
+ | LeViT | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 3.268 ms | 0 - 42 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
38
+ | LeViT | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.345 ms | 0 - 47 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
39
+ | LeViT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.099 ms | 0 - 92 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
40
+ | LeViT | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.547 ms | 0 - 42 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
41
+ | LeViT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 1.088 ms | 0 - 92 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
42
+ | LeViT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.635 ms | 0 - 44 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx) |
43
+ | LeViT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.777 ms | 0 - 52 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
44
+ | LeViT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.102 ms | 0 - 55 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx) |
45
+ | LeViT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 0.632 ms | 0 - 49 MB | NPU | [LeViT.tflite](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.tflite) |
46
+ | LeViT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.205 ms | 1 - 45 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx) |
47
+ | LeViT | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.686 ms | 17 - 17 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT.onnx) |
48
+ | LeViT | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 4.532 ms | 0 - 33 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
49
+ | LeViT | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 3.484 ms | 0 - 42 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
50
+ | LeViT | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 2.282 ms | 0 - 12 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
51
+ | LeViT | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 2.626 ms | 0 - 35 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
52
+ | LeViT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 2.279 ms | 0 - 12 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
53
+ | LeViT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 17.046 ms | 2 - 19 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx) |
54
+ | LeViT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.644 ms | 0 - 44 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
55
+ | LeViT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 13.766 ms | 2 - 33 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx) |
56
+ | LeViT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 1.802 ms | 0 - 36 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
57
+ | LeViT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 10.776 ms | 2 - 25 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx) |
58
+ | LeViT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 2.619 ms | 20 - 20 MB | NPU | [LeViT.dlc](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.dlc) |
59
+ | LeViT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 16.011 ms | 8 - 8 MB | NPU | [LeViT.onnx](https://huggingface.co/qualcomm/LeViT/blob/main/LeViT_w8a16.onnx) |
60
 
61
 
62
 
 
120
  LeViT
121
  Device : cs_8275 (ANDROID 14)
122
  Runtime : TFLITE
123
+ Estimated inference time (ms) : 3.3
124
+ Estimated peak memory usage (MB): [0, 42]
125
  Total # Ops : 306
126
  Compute Unit(s) : npu (306 ops) gpu (0 ops) cpu (0 ops)
127
  ```