Hugging Face's logo

qualcomm
/
EfficientViT-l2-cls

Image Classification
PyTorch
LiteRT
backbone
real_time
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
07edc10
·
verified ·
1 Parent(s): 5f4d068

v0.30.5

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.30.5 for changelog.

Files changed (2) hide show
  1. EfficientViT-l2-cls_w8a16.onnx +3 -0
  2. README.md +24 -23
EfficientViT-l2-cls_w8a16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbcd1ff0560a6a54520f8e6a154ae185c0a5fd6b40cfc236e5a538e750b299cf
3
+ size 255783504
README.md CHANGED
@@ -36,25 +36,26 @@ 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
- | EfficientViT-l2-cls | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 23.653 ms | 0 - 192 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
40
- | EfficientViT-l2-cls | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN | 24.101 ms | 1 - 10 MB | NPU | Use Export Script |
41
- | EfficientViT-l2-cls | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 13.836 ms | 0 - 203 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
42
- | EfficientViT-l2-cls | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN | 15.032 ms | 0 - 86 MB | NPU | Use Export Script |
43
- | EfficientViT-l2-cls | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 7.15 ms | 0 - 30 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
44
- | EfficientViT-l2-cls | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN | 7.213 ms | 1 - 3 MB | NPU | Use Export Script |
45
- | EfficientViT-l2-cls | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 9.065 ms | 0 - 192 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
46
- | EfficientViT-l2-cls | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN | 9.247 ms | 1 - 11 MB | NPU | Use Export Script |
47
- | EfficientViT-l2-cls | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 7.211 ms | 0 - 254 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
48
- | EfficientViT-l2-cls | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN | 7.306 ms | 0 - 21 MB | NPU | Use Export Script |
49
- | EfficientViT-l2-cls | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 29.603 ms | 0 - 397 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.onnx) |
50
- | EfficientViT-l2-cls | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 5.093 ms | 0 - 218 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
51
- | EfficientViT-l2-cls | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN | 5.275 ms | 1 - 106 MB | NPU | Use Export Script |
52
- | EfficientViT-l2-cls | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 20.317 ms | 1 - 162 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.onnx) |
53
- | EfficientViT-l2-cls | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 4.127 ms | 0 - 194 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
54
- | EfficientViT-l2-cls | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN | 4.022 ms | 1 - 81 MB | NPU | Use Export Script |
55
- | EfficientViT-l2-cls | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 18.276 ms | 1 - 126 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.onnx) |
56
- | EfficientViT-l2-cls | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 7.714 ms | 1 - 1 MB | NPU | Use Export Script |
57
- | EfficientViT-l2-cls | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 30.99 ms | 130 - 130 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.onnx) |
 
58
 
59
 
60
 
@@ -118,8 +119,8 @@ Profiling Results
118
  EfficientViT-l2-cls
119
  Device : cs_8275 (ANDROID 14)
120
  Runtime : TFLITE
121
- Estimated inference time (ms) : 23.7
122
- Estimated peak memory usage (MB): [0, 192]
123
  Total # Ops : 349
124
  Compute Unit(s) : npu (349 ops) gpu (0 ops) cpu (0 ops)
125
  ```
@@ -208,13 +209,13 @@ AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
208
  You can also run the demo on-device.
209
 
210
  ```bash
211
- python -m qai_hub_models.models.efficientvit_l2_cls.demo --on-device
212
  ```
213
 
214
  **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
215
  environment, please add the following to your cell (instead of the above).
216
  ```
217
- %run -m qai_hub_models.models.efficientvit_l2_cls.demo -- --on-device
218
  ```
219
 
220
 
 
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
+ | EfficientViT-l2-cls | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 23.448 ms | 0 - 204 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
40
+ | EfficientViT-l2-cls | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN | 23.771 ms | 1 - 10 MB | NPU | Use Export Script |
41
+ | EfficientViT-l2-cls | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 13.537 ms | 0 - 230 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
42
+ | EfficientViT-l2-cls | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN | 14.731 ms | 0 - 96 MB | NPU | Use Export Script |
43
+ | EfficientViT-l2-cls | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 7.107 ms | 0 - 216 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
44
+ | EfficientViT-l2-cls | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN | 7.219 ms | 1 - 3 MB | NPU | Use Export Script |
45
+ | EfficientViT-l2-cls | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 9.04 ms | 0 - 223 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
46
+ | EfficientViT-l2-cls | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN | 9.269 ms | 1 - 11 MB | NPU | Use Export Script |
47
+ | EfficientViT-l2-cls | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 7.141 ms | 0 - 114 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
48
+ | EfficientViT-l2-cls | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN | 7.113 ms | 0 - 24 MB | NPU | Use Export Script |
49
+ | EfficientViT-l2-cls | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 28.37 ms | 0 - 308 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.onnx) |
50
+ | EfficientViT-l2-cls | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 5.047 ms | 0 - 254 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
51
+ | EfficientViT-l2-cls | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN | 5.153 ms | 1 - 120 MB | NPU | Use Export Script |
52
+ | EfficientViT-l2-cls | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 19.859 ms | 1 - 170 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.onnx) |
53
+ | EfficientViT-l2-cls | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 4.638 ms | 0 - 227 MB | NPU | [EfficientViT-l2-cls.tflite](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.tflite) |
54
+ | EfficientViT-l2-cls | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN | 4.524 ms | 1 - 177 MB | NPU | Use Export Script |
55
+ | EfficientViT-l2-cls | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 17.415 ms | 1 - 139 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.onnx) |
56
+ | EfficientViT-l2-cls | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 7.59 ms | 1 - 1 MB | NPU | Use Export Script |
57
+ | EfficientViT-l2-cls | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 30.197 ms | 130 - 130 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls.onnx) |
58
+ | EfficientViT-l2-cls | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 241.313 ms | 100 - 100 MB | NPU | [EfficientViT-l2-cls.onnx](https://huggingface.co/qualcomm/EfficientViT-l2-cls/blob/main/EfficientViT-l2-cls_w8a16.onnx) |
59
 
60
 
61
 
 
119
  EfficientViT-l2-cls
120
  Device : cs_8275 (ANDROID 14)
121
  Runtime : TFLITE
122
+ Estimated inference time (ms) : 23.4
123
+ Estimated peak memory usage (MB): [0, 204]
124
  Total # Ops : 349
125
  Compute Unit(s) : npu (349 ops) gpu (0 ops) cpu (0 ops)
126
  ```
 
209
  You can also run the demo on-device.
210
 
211
  ```bash
212
+ python -m qai_hub_models.models.efficientvit_l2_cls.demo --eval-mode on-device
213
  ```
214
 
215
  **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
216
  environment, please add the following to your cell (instead of the above).
217
  ```
218
+ %run -m qai_hub_models.models.efficientvit_l2_cls.demo -- --eval-mode on-device
219
  ```
220
 
221