Hugging Face's logo

qualcomm
/
HRNetPose

Keypoint Detection
PyTorch
TF Lite
ONNX
android
Model card Files Community
bhushans commited on
Commit
00f9c33
·
verified ·
1 Parent(s): cbf4508

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +42 -25
README.md CHANGED
@@ -34,29 +34,30 @@ More details on model performance across various devices, can be found
34
 
35
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
- | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 2.811 ms | 0 - 2 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
38
- | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 2.954 ms | 1 - 17 MB | FP16 | NPU | [HRNetPose.so](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.so) |
39
- | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 2.892 ms | 0 - 58 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
40
- | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 2.082 ms | 0 - 124 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
41
- | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 2.144 ms | 1 - 35 MB | FP16 | NPU | [HRNetPose.so](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.so) |
42
- | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 2.167 ms | 1 - 150 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
43
- | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.969 ms | 0 - 59 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
44
- | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 2.11 ms | 0 - 32 MB | FP16 | NPU | Use Export Script |
45
- | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 2.16 ms | 0 - 71 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
46
- | HRNetPose | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 2.814 ms | 0 - 2 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
47
- | HRNetPose | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 2.714 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
48
- | HRNetPose | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 2.841 ms | 0 - 2 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
49
- | HRNetPose | SA8255 (Proxy) | SA8255P Proxy | QNN | 2.754 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
50
- | HRNetPose | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 2.806 ms | 0 - 2 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
51
- | HRNetPose | SA8775 (Proxy) | SA8775P Proxy | QNN | 2.752 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
52
- | HRNetPose | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 2.813 ms | 0 - 2 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
53
- | HRNetPose | SA8650 (Proxy) | SA8650P Proxy | QNN | 2.744 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
54
  | HRNetPose | SA8295P ADP | SA8295P | TFLITE | 4.653 ms | 0 - 51 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
55
- | HRNetPose | SA8295P ADP | SA8295P | QNN | 4.68 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
56
- | HRNetPose | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 3.726 ms | 0 - 110 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
57
- | HRNetPose | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 3.921 ms | 0 - 30 MB | FP16 | NPU | Use Export Script |
58
- | HRNetPose | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 2.992 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
59
- | HRNetPose | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.98 ms | 57 - 57 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
 
 
 
 
60
 
61
 
62
 
@@ -121,8 +122,8 @@ Profiling Results
121
  HRNetPose
122
  Device : Samsung Galaxy S23 (13)
123
  Runtime : TFLITE
124
- Estimated inference time (ms) : 2.8
125
- Estimated peak memory usage (MB): [0, 2]
126
  Total # Ops : 516
127
  Compute Unit(s) : NPU (516 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.hrnet_pose import
147
 
148
  # Load the model
 
149
 
150
  # Device
151
  device = hub.Device("Samsung Galaxy S23")
152
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
153
 
154
  ```
155
 
 
34
 
35
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
+ | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 2.853 ms | 0 - 59 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
38
+ | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 2.959 ms | 0 - 26 MB | FP16 | NPU | [HRNetPose.so](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.so) |
39
+ | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 2.914 ms | 0 - 595 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
40
+ | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 2.051 ms | 0 - 37 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
41
+ | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 2.176 ms | 0 - 31 MB | FP16 | NPU | [HRNetPose.so](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.so) |
42
+ | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 2.2 ms | 1 - 151 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
43
+ | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.689 ms | 0 - 33 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
44
+ | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 2.036 ms | 0 - 32 MB | FP16 | NPU | Use Export Script |
45
+ | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 2.141 ms | 0 - 72 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
46
+ | HRNetPose | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 2.818 ms | 0 - 49 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
47
+ | HRNetPose | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 2.706 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
48
+ | HRNetPose | SA7255P ADP | SA7255P | QNN | 103.12 ms | 1 - 7 MB | FP16 | NPU | Use Export Script |
49
+ | HRNetPose | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 2.834 ms | 0 - 39 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
50
+ | HRNetPose | SA8255 (Proxy) | SA8255P Proxy | QNN | 2.743 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
 
 
 
51
  | HRNetPose | SA8295P ADP | SA8295P | TFLITE | 4.653 ms | 0 - 51 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
52
+ | HRNetPose | SA8295P ADP | SA8295P | QNN | 5.172 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
53
+ | HRNetPose | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 2.855 ms | 0 - 58 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
54
+ | HRNetPose | SA8650 (Proxy) | SA8650P Proxy | QNN | 2.713 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
55
+ | HRNetPose | SA8775P ADP | SA8775P | TFLITE | 5.449 ms | 0 - 33 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
56
+ | HRNetPose | SA8775P ADP | SA8775P | QNN | 5.458 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
57
+ | HRNetPose | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 3.8 ms | 0 - 30 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
58
+ | HRNetPose | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 3.8 ms | 1 - 29 MB | FP16 | NPU | Use Export Script |
59
+ | HRNetPose | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 3.023 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
60
+ | HRNetPose | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.04 ms | 56 - 56 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
61
 
62
 
63
 
 
122
  HRNetPose
123
  Device : Samsung Galaxy S23 (13)
124
  Runtime : TFLITE
125
+ Estimated inference time (ms) : 2.9
126
+ Estimated peak memory usage (MB): [0, 59]
127
  Total # Ops : 516
128
  Compute Unit(s) : NPU (516 ops)
129
  ```
 
144
  import torch
145
 
146
  import qai_hub as hub
147
+ from qai_hub_models.models.hrnet_pose 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