Hugging Face's logo

qualcomm
/
WideResNet50

Image Classification
PyTorch
LiteRT
backbone
android
Model card Files
xet
Community
bhushans commited on
Commit
6e8076a
·
verified ·
1 Parent(s): f6a879b

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +41 -23
README.md CHANGED
@@ -35,29 +35,31 @@ More details on model performance across various devices, can be found
35
 
36
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
- | WideResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 4.873 ms | 0 - 2 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
39
- | WideResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 5.699 ms | 1 - 356 MB | FP16 | NPU | [WideResNet50.so](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.so) |
40
- | WideResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 5.221 ms | 0 - 161 MB | FP16 | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
41
- | WideResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 3.583 ms | 0 - 103 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
42
- | WideResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 4.182 ms | 1 - 29 MB | FP16 | NPU | [WideResNet50.so](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.so) |
43
- | WideResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 3.841 ms | 1 - 107 MB | FP16 | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
44
  | WideResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 3.41 ms | 0 - 32 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
45
- | WideResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 4.1 ms | 0 - 26 MB | FP16 | NPU | Use Export Script |
46
- | WideResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 3.601 ms | 0 - 36 MB | FP16 | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
47
- | WideResNet50 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 4.859 ms | 0 - 2 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
48
- | WideResNet50 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 5.016 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
49
- | WideResNet50 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 4.869 ms | 0 - 2 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
50
- | WideResNet50 | SA8255 (Proxy) | SA8255P Proxy | QNN | 5.04 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
51
- | WideResNet50 | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 4.863 ms | 0 - 2 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
52
- | WideResNet50 | SA8775 (Proxy) | SA8775P Proxy | QNN | 5.039 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
53
- | WideResNet50 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 4.873 ms | 0 - 2 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
54
- | WideResNet50 | SA8650 (Proxy) | SA8650P Proxy | QNN | 5.042 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
55
  | WideResNet50 | SA8295P ADP | SA8295P | TFLITE | 8.247 ms | 0 - 22 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
56
- | WideResNet50 | SA8295P ADP | SA8295P | QNN | 8.064 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
57
- | WideResNet50 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 7.135 ms | 0 - 91 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
58
- | WideResNet50 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 7.314 ms | 1 - 23 MB | FP16 | NPU | Use Export Script |
59
- | WideResNet50 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 4.898 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
60
- | WideResNet50 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 4.673 ms | 133 - 133 MB | FP16 | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
 
 
 
 
61
 
62
 
63
 
@@ -122,7 +124,7 @@ WideResNet50
122
  Device : Samsung Galaxy S23 (13)
123
  Runtime : TFLITE
124
  Estimated inference time (ms) : 4.9
125
- Estimated peak memory usage (MB): [0, 2]
126
  Total # Ops : 79
127
  Compute Unit(s) : NPU (79 ops)
128
  ```
@@ -143,13 +145,29 @@ in memory using the `jit.trace` and then call the `submit_compile_job` API.
143
  import torch
144
 
145
  import qai_hub as hub
146
- from qai_hub_models.models.wideresnet50 import
147
 
148
  # Load the model
 
149
 
150
  # Device
151
  device = hub.Device("Samsung Galaxy S23")
152
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
153
 
154
  ```
155
 
 
35
 
36
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
+ | WideResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 4.884 ms | 0 - 587 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
39
+ | WideResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 5.693 ms | 1 - 366 MB | FP16 | NPU | [WideResNet50.so](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.so) |
40
+ | WideResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 5.227 ms | 0 - 162 MB | FP16 | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
41
+ | WideResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 3.66 ms | 0 - 29 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
42
+ | WideResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 4.152 ms | 1 - 29 MB | FP16 | NPU | [WideResNet50.so](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.so) |
43
+ | WideResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 3.763 ms | 0 - 109 MB | FP16 | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
44
  | WideResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 3.41 ms | 0 - 32 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
45
+ | WideResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 4.012 ms | 1 - 27 MB | FP16 | NPU | Use Export Script |
46
+ | WideResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 3.552 ms | 0 - 39 MB | FP16 | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
47
+ | WideResNet50 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 4.849 ms | 0 - 587 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
48
+ | WideResNet50 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 5.029 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
49
+ | WideResNet50 | SA7255P ADP | SA7255P | TFLITE | 180.896 ms | 0 - 26 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
50
+ | WideResNet50 | SA7255P ADP | SA7255P | QNN | 184.744 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
51
+ | WideResNet50 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 4.881 ms | 0 - 567 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
52
+ | WideResNet50 | SA8255 (Proxy) | SA8255P Proxy | QNN | 5.039 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
 
 
53
  | WideResNet50 | SA8295P ADP | SA8295P | TFLITE | 8.247 ms | 0 - 22 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
54
+ | WideResNet50 | SA8295P ADP | SA8295P | QNN | 8.059 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
55
+ | WideResNet50 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 4.868 ms | 0 - 567 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
56
+ | WideResNet50 | SA8650 (Proxy) | SA8650P Proxy | QNN | 5.04 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
57
+ | WideResNet50 | SA8775P ADP | SA8775P | TFLITE | 8.492 ms | 0 - 25 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
58
+ | WideResNet50 | SA8775P ADP | SA8775P | QNN | 8.662 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
59
+ | WideResNet50 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 7.097 ms | 0 - 25 MB | FP16 | NPU | [WideResNet50.tflite](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.tflite) |
60
+ | WideResNet50 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 7.293 ms | 1 - 22 MB | FP16 | NPU | Use Export Script |
61
+ | WideResNet50 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 4.899 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
62
+ | WideResNet50 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 4.632 ms | 133 - 133 MB | FP16 | NPU | [WideResNet50.onnx](https://huggingface.co/qualcomm/WideResNet50/blob/main/WideResNet50.onnx) |
63
 
64
 
65
 
 
124
  Device : Samsung Galaxy S23 (13)
125
  Runtime : TFLITE
126
  Estimated inference time (ms) : 4.9
127
+ Estimated peak memory usage (MB): [0, 587]
128
  Total # Ops : 79
129
  Compute Unit(s) : NPU (79 ops)
130
  ```
 
145
  import torch
146
 
147
  import qai_hub as hub
148
+ from qai_hub_models.models.wideresnet50 import Model
149
 
150
  # Load the model
151
+ torch_model = Model.from_pretrained()
152
 
153
  # Device
154
  device = hub.Device("Samsung Galaxy S23")
155
 
156
+ # Trace model
157
+ input_shape = torch_model.get_input_spec()
158
+ sample_inputs = torch_model.sample_inputs()
159
+
160
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
161
+
162
+ # Compile model on a specific device
163
+ compile_job = hub.submit_compile_job(
164
+ model=pt_model,
165
+ device=device,
166
+ input_specs=torch_model.get_input_spec(),
167
+ )
168
+
169
+ # Get target model to run on-device
170
+ target_model = compile_job.get_target_model()
171
 
172
  ```
173