Hugging Face's logo

qualcomm
/
EfficientNet-B0

Image Classification
PyTorch
LiteRT
backbone
android
Model card Files Community
qaihm-bot commited on
Commit
d7a2460
·
verified ·
1 Parent(s): 49780ca

v0.34.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.34.0 for changelog.

Files changed (7) hide show
  1. EfficientNet-B0_w8a16.dlc +1 -1
  2. README.md +26 -35
  3. precompiled/qualcomm-qcs6490-proxy/sdk_versions.yml +3 -0
  4. precompiled/qualcomm-snapdragon-x-elite/EfficientNet-B0.bin +1 -1
  5. precompiled/qualcomm-snapdragon-x-elite/EfficientNet-B0.onnx.zip +2 -2
  6. precompiled/qualcomm-snapdragon-x-elite/EfficientNet-B0_w8a16.onnx.zip +2 -2
  7. precompiled/qualcomm-snapdragon-x-elite/sdk_versions.yml +5 -0
EfficientNet-B0_w8a16.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:12b448af968e654178034a2e59bb6a5f07a3748954c8d3da4076270d9342bad5
3
  size 7331255
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1cab3bcfb4b8055bcb87e80cd7bc420175dc1d533b8c307c7a8f26e80860df8b
3
  size 7331255
README.md CHANGED
@@ -24,6 +24,7 @@ More details on model performance across various devices, can be found
24
  [here](https://aihub.qualcomm.com/models/efficientnet_b0).
25
 
26
 
 
27
  ### Model Details
28
 
29
  - **Model Type:** Model_use_case.image_classification
@@ -36,52 +37,52 @@ 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
- | EfficientNet-B0 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 4.846 ms | 0 - 26 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
40
  | EfficientNet-B0 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 4.778 ms | 1 - 20 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
41
- | EfficientNet-B0 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 2.924 ms | 0 - 43 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
42
  | EfficientNet-B0 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 3.433 ms | 1 - 33 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
43
- | EfficientNet-B0 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.462 ms | 0 - 118 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
44
  | EfficientNet-B0 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.475 ms | 0 - 79 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
45
- | EfficientNet-B0 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.972 ms | 0 - 27 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
46
  | EfficientNet-B0 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.922 ms | 1 - 21 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
47
- | EfficientNet-B0 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 4.846 ms | 0 - 26 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
48
  | EfficientNet-B0 | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 4.778 ms | 1 - 20 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
49
  | EfficientNet-B0 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 1.466 ms | 0 - 118 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
50
  | EfficientNet-B0 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.468 ms | 0 - 93 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
51
- | EfficientNet-B0 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 3.6 ms | 0 - 30 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
52
  | EfficientNet-B0 | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 3.502 ms | 1 - 27 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
53
- | EfficientNet-B0 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 1.471 ms | 0 - 118 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
54
  | EfficientNet-B0 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.468 ms | 0 - 88 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
55
- | EfficientNet-B0 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 1.972 ms | 0 - 27 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
56
  | EfficientNet-B0 | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.922 ms | 1 - 21 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
57
  | EfficientNet-B0 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 1.464 ms | 0 - 118 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
58
  | EfficientNet-B0 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 1.468 ms | 0 - 88 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
59
  | EfficientNet-B0 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.36 ms | 0 - 94 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.onnx) |
60
- | EfficientNet-B0 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.048 ms | 0 - 42 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
61
  | EfficientNet-B0 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.023 ms | 1 - 37 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
62
  | EfficientNet-B0 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 0.957 ms | 0 - 36 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.onnx) |
63
- | EfficientNet-B0 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 0.838 ms | 0 - 27 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
64
  | EfficientNet-B0 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 0.993 ms | 1 - 26 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
65
  | EfficientNet-B0 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.013 ms | 1 - 25 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.onnx) |
66
  | EfficientNet-B0 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.946 ms | 79 - 79 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
67
  | EfficientNet-B0 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.459 ms | 13 - 13 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.onnx) |
68
- | EfficientNet-B0 | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 3.258 ms | 0 - 23 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
69
- | EfficientNet-B0 | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.95 ms | 0 - 38 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
70
- | EfficientNet-B0 | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.683 ms | 0 - 39 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
71
- | EfficientNet-B0 | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.973 ms | 0 - 25 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
72
- | EfficientNet-B0 | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 5.001 ms | 0 - 31 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
73
- | EfficientNet-B0 | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 3.258 ms | 0 - 23 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
74
- | EfficientNet-B0 | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.679 ms | 0 - 40 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
75
- | EfficientNet-B0 | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 2.364 ms | 0 - 26 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
76
- | EfficientNet-B0 | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.672 ms | 0 - 39 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
77
- | EfficientNet-B0 | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.973 ms | 0 - 25 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
78
- | EfficientNet-B0 | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 1.676 ms | 0 - 39 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
79
  | EfficientNet-B0 | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.676 ms | 0 - 61 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.onnx) |
80
- | EfficientNet-B0 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.128 ms | 0 - 41 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
81
  | EfficientNet-B0 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.071 ms | 0 - 46 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.onnx) |
82
- | EfficientNet-B0 | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 0.958 ms | 0 - 29 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
83
  | EfficientNet-B0 | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.024 ms | 0 - 36 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.onnx) |
84
- | EfficientNet-B0 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.95 ms | 28 - 28 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
85
  | EfficientNet-B0 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.746 ms | 8 - 8 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.onnx) |
86
 
87
 
@@ -140,17 +141,7 @@ device. This script does the following:
140
  ```bash
141
  python -m qai_hub_models.models.efficientnet_b0.export
142
  ```
143
- ```
144
- Profiling Results
145
- ------------------------------------------------------------
146
- EfficientNet-B0
147
- Device : cs_8275 (ANDROID 14)
148
- Runtime : TFLITE
149
- Estimated inference time (ms) : 4.8
150
- Estimated peak memory usage (MB): [0, 26]
151
- Total # Ops : 245
152
- Compute Unit(s) : npu (245 ops) gpu (0 ops) cpu (0 ops)
153
- ```
154
 
155
 
156
  ## How does this work?
 
24
  [here](https://aihub.qualcomm.com/models/efficientnet_b0).
25
 
26
 
27
+
28
  ### Model Details
29
 
30
  - **Model Type:** Model_use_case.image_classification
 
37
 
38
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
39
  |---|---|---|---|---|---|---|---|---|
40
+ | EfficientNet-B0 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 4.903 ms | 0 - 26 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
41
  | EfficientNet-B0 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 4.778 ms | 1 - 20 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
42
+ | EfficientNet-B0 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 2.938 ms | 0 - 40 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
43
  | EfficientNet-B0 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 3.433 ms | 1 - 33 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
44
+ | EfficientNet-B0 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.46 ms | 0 - 118 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
45
  | EfficientNet-B0 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.475 ms | 0 - 79 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
46
+ | EfficientNet-B0 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.973 ms | 0 - 28 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
47
  | EfficientNet-B0 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.922 ms | 1 - 21 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
48
+ | EfficientNet-B0 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 4.903 ms | 0 - 26 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
49
  | EfficientNet-B0 | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 4.778 ms | 1 - 20 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
50
  | EfficientNet-B0 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 1.466 ms | 0 - 118 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
51
  | EfficientNet-B0 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.468 ms | 0 - 93 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
52
+ | EfficientNet-B0 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 3.556 ms | 0 - 33 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
53
  | EfficientNet-B0 | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 3.502 ms | 1 - 27 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
54
+ | EfficientNet-B0 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 1.468 ms | 0 - 118 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
55
  | EfficientNet-B0 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.468 ms | 0 - 88 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
56
+ | EfficientNet-B0 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 1.973 ms | 0 - 28 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
57
  | EfficientNet-B0 | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.922 ms | 1 - 21 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
58
  | EfficientNet-B0 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 1.464 ms | 0 - 118 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
59
  | EfficientNet-B0 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 1.468 ms | 0 - 88 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
60
  | EfficientNet-B0 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.36 ms | 0 - 94 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.onnx) |
61
+ | EfficientNet-B0 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.047 ms | 0 - 43 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
62
  | EfficientNet-B0 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.023 ms | 1 - 37 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
63
  | EfficientNet-B0 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 0.957 ms | 0 - 36 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.onnx) |
64
+ | EfficientNet-B0 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 0.837 ms | 0 - 32 MB | NPU | [EfficientNet-B0.tflite](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.tflite) |
65
  | EfficientNet-B0 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 0.993 ms | 1 - 26 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
66
  | EfficientNet-B0 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.013 ms | 1 - 25 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.onnx) |
67
  | EfficientNet-B0 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.946 ms | 79 - 79 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.dlc) |
68
  | EfficientNet-B0 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.459 ms | 13 - 13 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0.onnx) |
69
+ | EfficientNet-B0 | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 3.241 ms | 0 - 23 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
70
+ | EfficientNet-B0 | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.949 ms | 0 - 39 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
71
+ | EfficientNet-B0 | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.67 ms | 0 - 39 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
72
+ | EfficientNet-B0 | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.991 ms | 0 - 25 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
73
+ | EfficientNet-B0 | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 4.995 ms | 0 - 31 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
74
+ | EfficientNet-B0 | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 3.241 ms | 0 - 23 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
75
+ | EfficientNet-B0 | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.681 ms | 0 - 39 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
76
+ | EfficientNet-B0 | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 2.324 ms | 0 - 28 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
77
+ | EfficientNet-B0 | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.685 ms | 0 - 39 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
78
+ | EfficientNet-B0 | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.991 ms | 0 - 25 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
79
+ | EfficientNet-B0 | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 1.683 ms | 0 - 39 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
80
  | EfficientNet-B0 | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.676 ms | 0 - 61 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.onnx) |
81
+ | EfficientNet-B0 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.128 ms | 0 - 40 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
82
  | EfficientNet-B0 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.071 ms | 0 - 46 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.onnx) |
83
+ | EfficientNet-B0 | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 0.951 ms | 0 - 30 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
84
  | EfficientNet-B0 | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.024 ms | 0 - 36 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.onnx) |
85
+ | EfficientNet-B0 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.942 ms | 28 - 28 MB | NPU | [EfficientNet-B0.dlc](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.dlc) |
86
  | EfficientNet-B0 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.746 ms | 8 - 8 MB | NPU | [EfficientNet-B0.onnx](https://huggingface.co/qualcomm/EfficientNet-B0/blob/main/EfficientNet-B0_w8a16.onnx) |
87
 
88
 
 
141
  ```bash
142
  python -m qai_hub_models.models.efficientnet_b0.export
143
  ```
144
+
 
 
 
 
 
 
 
 
 
 
145
 
146
 
147
  ## How does this work?
precompiled/qualcomm-qcs6490-proxy/sdk_versions.yml ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ sdk_versions:
2
+ qnn_context_binary:
3
+ qairt: 2.34.2.250528164111_119506
precompiled/qualcomm-snapdragon-x-elite/EfficientNet-B0.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:11d5976daab404de1eac4a3402788897871099fbf389c2bbbeec647866f0f42d
3
  size 14141752
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:256c1b81465397cda02e56a456caf1f5992ee52a144cb7dd581f90e73edfaa83
3
  size 14141752
precompiled/qualcomm-snapdragon-x-elite/EfficientNet-B0.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5b5cff6f40057df95b7dec3bcf4b4771c0e5de7a6dbd9244cbca424de372095f
3
- size 10392416
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:067a7490ff5b011399052c2f305f3b4b715595679e2a966037785dfdf96639d8
3
+ size 10392412
precompiled/qualcomm-snapdragon-x-elite/EfficientNet-B0_w8a16.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f682ee1c15ca41ce59134e6d8f25dfb848cdcacef7102e4dd85aa7fe927e1b52
3
- size 5176091
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:221bd0d5e7ccb86c43ee03d348489fd92961b0b12a3c07be51522171800bac7e
3
+ size 5176095
precompiled/qualcomm-snapdragon-x-elite/sdk_versions.yml ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ sdk_versions:
2
+ qnn_context_binary:
3
+ qairt: 2.34.2.250528164111_119506
4
+ precompiled_qnn_onnx:
5
+ qairt: 2.33.2.250410134701_117956