Hugging Face's logo

qualcomm
/
EfficientNet-B4

Image Classification
PyTorch
LiteRT
backbone
bu_auto
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
0ac31c8
·
verified ·
1 Parent(s): 8427821

v0.32.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.32.0 for changelog.

Files changed (6) hide show
  1. .gitattributes +1 -0
  2. DEPLOYMENT_MODEL_LICENSE.pdf +3 -0
  3. EfficientNet-B4.onnx +2 -2
  4. EfficientNet-B4_w8a16.dlc +2 -2
  5. LICENSE +2 -0
  6. README.md +16 -15
.gitattributes CHANGED
@@ -37,3 +37,4 @@ EfficientNet-B4.so filter=lfs diff=lfs merge=lfs -text
37
  EfficientNet-B4_w8a16.so filter=lfs diff=lfs merge=lfs -text
38
  EfficientNet-B4.dlc filter=lfs diff=lfs merge=lfs -text
39
  EfficientNet-B4_w8a16.dlc filter=lfs diff=lfs merge=lfs -text
 
 
37
  EfficientNet-B4_w8a16.so filter=lfs diff=lfs merge=lfs -text
38
  EfficientNet-B4.dlc filter=lfs diff=lfs merge=lfs -text
39
  EfficientNet-B4_w8a16.dlc filter=lfs diff=lfs merge=lfs -text
40
+ DEPLOYMENT_MODEL_LICENSE.pdf filter=lfs diff=lfs merge=lfs -text
DEPLOYMENT_MODEL_LICENSE.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4409f93b0e82531303b3e10f52f1fdfb56467a25f05b7441c6bbd8bb8a64b42c
3
+ size 109629
EfficientNet-B4.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4b4e414a2c352ac70593f9b6f8bfe145b973a40917658925f31b5ddf40eb1b58
3
- size 77260178
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbf4759fe77334f86a52608f4a339930600aeb8fbd227116b4ec24c7500b17a5
3
+ size 77274917
EfficientNet-B4_w8a16.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a971af24f84a3782026f1a1c6ecd8700d911f2dd74f686a301d2ed9acc82b8dd
3
- size 25206152
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08f3d28aa7ecb433a0274ea50ff97fa579498fd010aca74650bfde674708db4f
3
+ size 25206095
LICENSE ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ The license of the original trained model can be found at https://github.com/pytorch/vision/blob/main/LICENSE.
2
+ The license for the deployable model files (.tflite, .onnx, .dlc, .bin, etc.) can be found in DEPLOYMENT_MODEL_LICENSE.pdf.
README.md CHANGED
@@ -30,8 +30,9 @@ More details on model performance across various devices, can be found
30
  - **Model Stats:**
31
  - Model checkpoint: Imagenet
32
  - Input resolution: 380x380
33
- - Number of parameters: 19.34M
34
- - Model size: 74.5 MB
 
35
 
36
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
@@ -45,24 +46,24 @@ More details on model performance across various devices, can be found
45
  | EfficientNet-B4 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 4.212 ms | 1 - 33 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
46
  | EfficientNet-B4 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 3.314 ms | 0 - 421 MB | NPU | [EfficientNet-B4.tflite](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.tflite) |
47
  | EfficientNet-B4 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 3.183 ms | 0 - 64 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
48
- | EfficientNet-B4 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 3.15 ms | 0 - 104 MB | NPU | [EfficientNet-B4.onnx](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.onnx) |
49
  | EfficientNet-B4 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 2.447 ms | 0 - 84 MB | NPU | [EfficientNet-B4.tflite](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.tflite) |
50
  | EfficientNet-B4 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 2.356 ms | 1 - 47 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
51
- | EfficientNet-B4 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.388 ms | 0 - 54 MB | NPU | [EfficientNet-B4.onnx](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.onnx) |
52
  | EfficientNet-B4 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 2.333 ms | 0 - 71 MB | NPU | [EfficientNet-B4.tflite](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.tflite) |
53
  | EfficientNet-B4 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 2.176 ms | 1 - 38 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
54
- | EfficientNet-B4 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.963 ms | 0 - 40 MB | NPU | [EfficientNet-B4.onnx](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.onnx) |
55
  | EfficientNet-B4 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 3.896 ms | 245 - 245 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
56
- | EfficientNet-B4 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.431 ms | 45 - 45 MB | NPU | [EfficientNet-B4.onnx](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.onnx) |
57
- | EfficientNet-B4 | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 6.541 ms | 0 - 56 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
58
- | EfficientNet-B4 | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 4.195 ms | 0 - 71 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
59
- | EfficientNet-B4 | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 3.403 ms | 0 - 19 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
60
- | EfficientNet-B4 | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 3.86 ms | 0 - 57 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
61
- | EfficientNet-B4 | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 11.855 ms | 0 - 73 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
62
- | EfficientNet-B4 | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 3.416 ms | 0 - 18 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
63
- | EfficientNet-B4 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 2.288 ms | 0 - 78 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
64
- | EfficientNet-B4 | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 1.929 ms | 0 - 61 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
65
- | EfficientNet-B4 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 3.872 ms | 104 - 104 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
66
 
67
 
68
 
 
30
  - **Model Stats:**
31
  - Model checkpoint: Imagenet
32
  - Input resolution: 380x380
33
+ - Number of parameters: 19.3M
34
+ - Model size (float): 73.6 MB
35
+ - Model size (w8a16): 24.0 MB
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
 
46
  | EfficientNet-B4 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 4.212 ms | 1 - 33 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
47
  | EfficientNet-B4 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 3.314 ms | 0 - 421 MB | NPU | [EfficientNet-B4.tflite](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.tflite) |
48
  | EfficientNet-B4 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 3.183 ms | 0 - 64 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
49
+ | EfficientNet-B4 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 3.176 ms | 0 - 107 MB | NPU | [EfficientNet-B4.onnx](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.onnx) |
50
  | EfficientNet-B4 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 2.447 ms | 0 - 84 MB | NPU | [EfficientNet-B4.tflite](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.tflite) |
51
  | EfficientNet-B4 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 2.356 ms | 1 - 47 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
52
+ | EfficientNet-B4 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.371 ms | 0 - 54 MB | NPU | [EfficientNet-B4.onnx](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.onnx) |
53
  | EfficientNet-B4 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 2.333 ms | 0 - 71 MB | NPU | [EfficientNet-B4.tflite](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.tflite) |
54
  | EfficientNet-B4 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 2.176 ms | 1 - 38 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
55
+ | EfficientNet-B4 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 2.274 ms | 0 - 39 MB | NPU | [EfficientNet-B4.onnx](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.onnx) |
56
  | EfficientNet-B4 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 3.896 ms | 245 - 245 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.dlc) |
57
+ | EfficientNet-B4 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.42 ms | 46 - 46 MB | NPU | [EfficientNet-B4.onnx](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4.onnx) |
58
+ | EfficientNet-B4 | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 6.519 ms | 0 - 56 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
59
+ | EfficientNet-B4 | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 4.23 ms | 0 - 67 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
60
+ | EfficientNet-B4 | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 3.416 ms | 0 - 16 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
61
+ | EfficientNet-B4 | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 3.829 ms | 0 - 56 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
62
+ | EfficientNet-B4 | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 11.862 ms | 0 - 74 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
63
+ | EfficientNet-B4 | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 3.413 ms | 0 - 16 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
64
+ | EfficientNet-B4 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 2.278 ms | 0 - 80 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
65
+ | EfficientNet-B4 | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 1.939 ms | 0 - 61 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
66
+ | EfficientNet-B4 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 3.865 ms | 119 - 119 MB | NPU | [EfficientNet-B4.dlc](https://huggingface.co/qualcomm/EfficientNet-B4/blob/main/EfficientNet-B4_w8a16.dlc) |
67
 
68
 
69