Hugging Face's logo

qualcomm
/
HRNetPose

Keypoint Detection
PyTorch
LiteRT
ONNX
android
Model card Files Community
qaihm-bot commited on
Commit
9a72f1a
·
verified ·
1 Parent(s): 53c0322

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +33 -33
README.md CHANGED
@@ -1,6 +1,6 @@
1
  ---
2
  library_name: pytorch
3
- license: other
4
  tags:
5
  - android
6
  pipeline_tag: keypoint-detection
@@ -15,7 +15,7 @@ pipeline_tag: keypoint-detection
15
 
16
  HRNet performs pose estimation in high-resolution representations.
17
 
18
- This model is an implementation of HRNetPose found [here](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/hrnet_posenet).
19
 
20
 
21
  This repository provides scripts to run HRNetPose on Qualcomm® devices.
@@ -34,35 +34,35 @@ 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 | 2.674 ms | 0 - 22 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
38
- | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 2.677 ms | 1 - 3 MB | FP16 | NPU | [HRNetPose.so](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.so) |
39
- | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 2.74 ms | 0 - 166 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
40
- | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 1.959 ms | 0 - 117 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
41
- | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 1.999 ms | 0 - 19 MB | FP16 | NPU | [HRNetPose.so](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.so) |
42
- | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 2.01 ms | 0 - 95 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
43
- | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.853 ms | 0 - 75 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
44
  | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 1.851 ms | 0 - 39 MB | FP16 | NPU | Use Export Script |
45
- | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 1.988 ms | 0 - 52 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
46
- | HRNetPose | SA7255P ADP | SA7255P | TFLITE | 102.841 ms | 0 - 72 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
47
- | HRNetPose | SA7255P ADP | SA7255P | QNN | 102.695 ms | 1 - 10 MB | FP16 | NPU | Use Export Script |
48
- | HRNetPose | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 2.664 ms | 0 - 92 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
49
- | HRNetPose | SA8255 (Proxy) | SA8255P Proxy | QNN | 2.65 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
50
- | HRNetPose | SA8295P ADP | SA8295P | TFLITE | 4.582 ms | 0 - 67 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
51
- | HRNetPose | SA8295P ADP | SA8295P | QNN | 4.533 ms | 1 - 18 MB | FP16 | NPU | Use Export Script |
52
- | HRNetPose | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 2.662 ms | 0 - 25 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
53
- | HRNetPose | SA8650 (Proxy) | SA8650P Proxy | QNN | 2.693 ms | 1 - 3 MB | FP16 | NPU | Use Export Script |
54
- | HRNetPose | SA8775P ADP | SA8775P | TFLITE | 5.264 ms | 0 - 71 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
55
- | HRNetPose | SA8775P ADP | SA8775P | QNN | 5.164 ms | 1 - 10 MB | FP16 | NPU | Use Export Script |
56
- | HRNetPose | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 102.841 ms | 0 - 72 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
57
- | HRNetPose | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 102.695 ms | 1 - 10 MB | FP16 | NPU | Use Export Script |
58
- | HRNetPose | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 2.667 ms | 0 - 83 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
59
- | HRNetPose | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 2.65 ms | 1 - 3 MB | FP16 | NPU | Use Export Script |
60
- | HRNetPose | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 5.264 ms | 0 - 71 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
61
- | HRNetPose | QCS9075 (Proxy) | QCS9075 Proxy | QNN | 5.164 ms | 1 - 10 MB | FP16 | NPU | Use Export Script |
62
- | HRNetPose | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 3.925 ms | 0 - 118 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
63
- | HRNetPose | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 4.021 ms | 0 - 46 MB | FP16 | NPU | Use Export Script |
64
- | HRNetPose | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 2.901 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
65
- | HRNetPose | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.71 ms | 56 - 56 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
66
 
67
 
68
 
@@ -127,7 +127,7 @@ HRNetPose
127
  Device : Samsung Galaxy S23 (13)
128
  Runtime : TFLITE
129
  Estimated inference time (ms) : 2.7
130
- Estimated peak memory usage (MB): [0, 22]
131
  Total # Ops : 516
132
  Compute Unit(s) : NPU (516 ops)
133
  ```
@@ -247,14 +247,14 @@ Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
247
 
248
  ## License
249
  * The license for the original implementation of HRNetPose can be found
250
- [here](https://github.com/quic/aimet-model-zoo/blob/develop/LICENSE.pdf).
251
  * The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf)
252
 
253
 
254
 
255
  ## References
256
  * [Deep High-Resolution Representation Learning for Human Pose Estimation](https://arxiv.org/abs/1902.09212)
257
- * [Source Model Implementation](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/hrnet_posenet)
258
 
259
 
260
 
 
1
  ---
2
  library_name: pytorch
3
+ license: mit
4
  tags:
5
  - android
6
  pipeline_tag: keypoint-detection
 
15
 
16
  HRNet performs pose estimation in high-resolution representations.
17
 
18
+ This model is an implementation of HRNetPose found [here](https://github.com/leoxiaobin/deep-high-resolution-net.pytorch).
19
 
20
 
21
  This repository provides scripts to run HRNetPose on Qualcomm® devices.
 
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 | 2.661 ms | 0 - 20 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
38
+ | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 2.761 ms | 0 - 20 MB | FP16 | NPU | [HRNetPose.so](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.so) |
39
+ | HRNetPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 2.727 ms | 0 - 163 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
40
+ | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 1.942 ms | 0 - 115 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
41
+ | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 1.959 ms | 0 - 54 MB | FP16 | NPU | [HRNetPose.so](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.so) |
42
+ | HRNetPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 2.003 ms | 0 - 92 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
43
+ | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.852 ms | 0 - 74 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
44
  | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 1.851 ms | 0 - 39 MB | FP16 | NPU | Use Export Script |
45
+ | HRNetPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 1.97 ms | 0 - 52 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
46
+ | HRNetPose | SA7255P ADP | SA7255P | TFLITE | 102.856 ms | 0 - 72 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
47
+ | HRNetPose | SA7255P ADP | SA7255P | QNN | 102.654 ms | 1 - 10 MB | FP16 | NPU | Use Export Script |
48
+ | HRNetPose | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 2.66 ms | 0 - 19 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
49
+ | HRNetPose | SA8255 (Proxy) | SA8255P Proxy | QNN | 2.664 ms | 1 - 3 MB | FP16 | NPU | Use Export Script |
50
+ | HRNetPose | SA8295P ADP | SA8295P | TFLITE | 4.59 ms | 0 - 67 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
51
+ | HRNetPose | SA8295P ADP | SA8295P | QNN | 4.52 ms | 0 - 18 MB | FP16 | NPU | Use Export Script |
52
+ | HRNetPose | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 2.659 ms | 0 - 46 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
53
+ | HRNetPose | SA8650 (Proxy) | SA8650P Proxy | QNN | 2.692 ms | 1 - 4 MB | FP16 | NPU | Use Export Script |
54
+ | HRNetPose | SA8775P ADP | SA8775P | TFLITE | 5.313 ms | 0 - 71 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
55
+ | HRNetPose | SA8775P ADP | SA8775P | QNN | 5.176 ms | 1 - 11 MB | FP16 | NPU | Use Export Script |
56
+ | HRNetPose | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 102.856 ms | 0 - 72 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
57
+ | HRNetPose | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 102.654 ms | 1 - 10 MB | FP16 | NPU | Use Export Script |
58
+ | HRNetPose | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 2.672 ms | 0 - 148 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
59
+ | HRNetPose | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 2.658 ms | 1 - 3 MB | FP16 | NPU | Use Export Script |
60
+ | HRNetPose | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 5.313 ms | 0 - 71 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
61
+ | HRNetPose | QCS9075 (Proxy) | QCS9075 Proxy | QNN | 5.176 ms | 1 - 11 MB | FP16 | NPU | Use Export Script |
62
+ | HRNetPose | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 3.887 ms | 0 - 115 MB | FP16 | NPU | [HRNetPose.tflite](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.tflite) |
63
+ | HRNetPose | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 3.845 ms | 0 - 48 MB | FP16 | NPU | Use Export Script |
64
+ | HRNetPose | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 2.831 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
65
+ | HRNetPose | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.722 ms | 55 - 55 MB | FP16 | NPU | [HRNetPose.onnx](https://huggingface.co/qualcomm/HRNetPose/blob/main/HRNetPose.onnx) |
66
 
67
 
68
 
 
127
  Device : Samsung Galaxy S23 (13)
128
  Runtime : TFLITE
129
  Estimated inference time (ms) : 2.7
130
+ Estimated peak memory usage (MB): [0, 20]
131
  Total # Ops : 516
132
  Compute Unit(s) : NPU (516 ops)
133
  ```
 
247
 
248
  ## License
249
  * The license for the original implementation of HRNetPose can be found
250
+ [here](https://github.com/leoxiaobin/deep-high-resolution-net.pytorch/blob/master/LICENSE).
251
  * The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf)
252
 
253
 
254
 
255
  ## References
256
  * [Deep High-Resolution Representation Learning for Human Pose Estimation](https://arxiv.org/abs/1902.09212)
257
+ * [Source Model Implementation](https://github.com/leoxiaobin/deep-high-resolution-net.pytorch)
258
 
259
 
260