Hugging Face's logo

qualcomm
/
MNASNet05

Image Classification
PyTorch
LiteRT
ONNX
backbone
android
Model card Files Community
bhushans commited on
Commit
358eff4
·
verified ·
1 Parent(s): d28dfd6

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +40 -23
README.md CHANGED
@@ -35,29 +35,30 @@ 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
- | MNASNet05 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 0.753 ms | 0 - 1 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
39
- | MNASNet05 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 0.817 ms | 1 - 22 MB | FP16 | NPU | [MNASNet05.so](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.so) |
40
- | MNASNet05 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 0.746 ms | 1 - 145 MB | FP16 | NPU | [MNASNet05.onnx](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.onnx) |
41
- | MNASNet05 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 0.516 ms | 0 - 49 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
42
- | MNASNet05 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 0.56 ms | 0 - 16 MB | FP16 | NPU | [MNASNet05.so](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.so) |
43
- | MNASNet05 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 0.541 ms | 0 - 53 MB | FP16 | NPU | [MNASNet05.onnx](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.onnx) |
44
- | MNASNet05 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 0.506 ms | 0 - 22 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
45
- | MNASNet05 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 0.552 ms | 0 - 12 MB | FP16 | NPU | Use Export Script |
46
- | MNASNet05 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 0.467 ms | 0 - 23 MB | FP16 | NPU | [MNASNet05.onnx](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.onnx) |
47
- | MNASNet05 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 0.755 ms | 0 - 1 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
48
  | MNASNet05 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 0.76 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
49
- | MNASNet05 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 0.757 ms | 0 - 6 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
50
- | MNASNet05 | SA8255 (Proxy) | SA8255P Proxy | QNN | 0.758 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
51
- | MNASNet05 | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 0.753 ms | 0 - 2 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
52
- | MNASNet05 | SA8775 (Proxy) | SA8775P Proxy | QNN | 0.759 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
53
- | MNASNet05 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 0.757 ms | 0 - 153 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
 
 
54
  | MNASNet05 | SA8650 (Proxy) | SA8650P Proxy | QNN | 0.764 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
55
- | MNASNet05 | SA8295P ADP | SA8295P | TFLITE | 1.472 ms | 0 - 22 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
56
- | MNASNet05 | SA8295P ADP | SA8295P | QNN | 1.622 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
57
- | MNASNet05 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 1.024 ms | 0 - 51 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
58
- | MNASNet05 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 1.107 ms | 1 - 18 MB | FP16 | NPU | Use Export Script |
59
- | MNASNet05 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 0.904 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
60
- | MNASNet05 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 0.841 ms | 7 - 7 MB | FP16 | NPU | [MNASNet05.onnx](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.onnx) |
61
 
62
 
63
 
@@ -122,7 +123,7 @@ MNASNet05
122
  Device : Samsung Galaxy S23 (13)
123
  Runtime : TFLITE
124
  Estimated inference time (ms) : 0.8
125
- Estimated peak memory usage (MB): [0, 1]
126
  Total # Ops : 71
127
  Compute Unit(s) : NPU (71 ops)
128
  ```
@@ -143,13 +144,29 @@ 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.mnasnet05 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
+ | MNASNet05 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 0.759 ms | 0 - 42 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
39
+ | MNASNet05 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 0.824 ms | 0 - 164 MB | FP16 | NPU | [MNASNet05.so](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.so) |
40
+ | MNASNet05 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 0.771 ms | 0 - 2 MB | FP16 | NPU | [MNASNet05.onnx](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.onnx) |
41
+ | MNASNet05 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 0.51 ms | 0 - 10 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
42
+ | MNASNet05 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 0.563 ms | 0 - 13 MB | FP16 | NPU | [MNASNet05.so](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.so) |
43
+ | MNASNet05 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 0.555 ms | 0 - 53 MB | FP16 | NPU | [MNASNet05.onnx](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.onnx) |
44
+ | MNASNet05 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 0.512 ms | 0 - 11 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
45
+ | MNASNet05 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 0.555 ms | 0 - 12 MB | FP16 | NPU | Use Export Script |
46
+ | MNASNet05 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 0.564 ms | 0 - 23 MB | FP16 | NPU | [MNASNet05.onnx](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.onnx) |
47
+ | MNASNet05 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 0.757 ms | 0 - 55 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
48
  | MNASNet05 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 0.76 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
49
+ | MNASNet05 | SA7255P ADP | SA7255P | TFLITE | 10.829 ms | 0 - 12 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
50
+ | MNASNet05 | SA7255P ADP | SA7255P | QNN | 11.045 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
51
+ | MNASNet05 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 0.759 ms | 0 - 32 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
52
+ | MNASNet05 | SA8255 (Proxy) | SA8255P Proxy | QNN | 0.767 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
53
+ | MNASNet05 | SA8295P ADP | SA8295P | TFLITE | 1.468 ms | 0 - 10 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
54
+ | MNASNet05 | SA8295P ADP | SA8295P | QNN | 1.64 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
55
+ | MNASNet05 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 0.76 ms | 0 - 109 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
56
  | MNASNet05 | SA8650 (Proxy) | SA8650P Proxy | QNN | 0.764 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
57
+ | MNASNet05 | SA8775P ADP | SA8775P | TFLITE | 1.351 ms | 0 - 10 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
58
+ | MNASNet05 | SA8775P ADP | SA8775P | QNN | 1.504 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
59
+ | MNASNet05 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 1.039 ms | 0 - 16 MB | FP16 | NPU | [MNASNet05.tflite](https://huggingface.co/qualcomm/MNASNet05/blob/main/MNASNet05.tflite) |
60
+ | MNASNet05 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 1.107 ms | 1 - 15 MB | FP16 | NPU | Use Export Script |
61
+ | MNASNet05 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 0.902 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
 
62
 
63
 
64
 
 
123
  Device : Samsung Galaxy S23 (13)
124
  Runtime : TFLITE
125
  Estimated inference time (ms) : 0.8
126
+ Estimated peak memory usage (MB): [0, 42]
127
  Total # Ops : 71
128
  Compute Unit(s) : NPU (71 ops)
129
  ```
 
144
  import torch
145
 
146
  import qai_hub as hub
147
+ from qai_hub_models.models.mnasnet05 import Model
148
 
149
  # Load the model
150
+ torch_model = Model.from_pretrained()
151
 
152
  # Device
153
  device = hub.Device("Samsung Galaxy S23")
154
 
155
+ # Trace model
156
+ input_shape = torch_model.get_input_spec()
157
+ sample_inputs = torch_model.sample_inputs()
158
+
159
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
160
+
161
+ # Compile model on a specific device
162
+ compile_job = hub.submit_compile_job(
163
+ model=pt_model,
164
+ device=device,
165
+ input_specs=torch_model.get_input_spec(),
166
+ )
167
+
168
+ # Get target model to run on-device
169
+ target_model = compile_job.get_target_model()
170
 
171
  ```
172