Hugging Face's logo

qualcomm
/
LiteHRNet

Keypoint Detection
PyTorch
LiteRT
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
b5d5412
·
verified ·
1 Parent(s): 619c86a

v0.42.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.42.0 for changelog.

Files changed (2) hide show
  1. LiteHRNet_float.onnx.zip +1 -1
  2. README.md +20 -20
LiteHRNet_float.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3a1b2d0c36d20c3dcb5a9d5ef54fc6ac10c51666c41848d65f12b39ae01743de
3
  size 4220859
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8af57ac3945fb62dabece4dded2d477c3628b3398ff5d55bd69d5f5f04148f72
3
  size 4220859
README.md CHANGED
@@ -34,22 +34,22 @@ More details on model performance across various devices, can be found
34
 
35
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
- | LiteHRNet | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 8.916 ms | 0 - 59 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
38
- | LiteHRNet | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 5.264 ms | 0 - 67 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
39
- | LiteHRNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 4.456 ms | 0 - 19 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
40
- | LiteHRNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 5.555 ms | 0 - 28 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
41
- | LiteHRNet | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 5.436 ms | 0 - 59 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
42
- | LiteHRNet | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 8.916 ms | 0 - 59 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
43
- | LiteHRNet | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 4.448 ms | 0 - 18 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
44
- | LiteHRNet | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 6.488 ms | 0 - 63 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
45
- | LiteHRNet | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 4.471 ms | 0 - 19 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
46
- | LiteHRNet | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 5.436 ms | 0 - 59 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
47
- | LiteHRNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 2.766 ms | 0 - 66 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
48
- | LiteHRNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 3.401 ms | 0 - 60 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
49
- | LiteHRNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 2.335 ms | 0 - 59 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
50
- | LiteHRNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 2.92 ms | 0 - 50 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
51
- | LiteHRNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 2.12 ms | 0 - 59 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
52
- | LiteHRNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 2.784 ms | 1 - 53 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
53
  | LiteHRNet | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 5.849 ms | 4 - 4 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
54
 
55
 
@@ -66,9 +66,9 @@ pip install "qai-hub-models[litehrnet]"
66
  ```
67
 
68
 
69
- ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
70
 
71
- Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
72
  Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
73
 
74
  With this API token, you can configure your client to run models on the cloud
@@ -76,7 +76,7 @@ hosted devices.
76
  ```bash
77
  qai-hub configure --api_token API_TOKEN
78
  ```
79
- Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information.
80
 
81
 
82
 
@@ -187,7 +187,7 @@ With the output of the model, you can compute like PSNR, relative errors or
187
  spot check the output with expected output.
188
 
189
  **Note**: This on-device profiling and inference requires access to Qualcomm®
190
- AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
191
 
192
 
193
 
 
34
 
35
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
+ | LiteHRNet | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 8.793 ms | 0 - 60 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
38
+ | LiteHRNet | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 5.265 ms | 0 - 68 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
39
+ | LiteHRNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 4.47 ms | 0 - 20 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
40
+ | LiteHRNet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 5.592 ms | 0 - 27 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
41
+ | LiteHRNet | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 5.427 ms | 0 - 60 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
42
+ | LiteHRNet | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 8.793 ms | 0 - 60 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
43
+ | LiteHRNet | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 4.478 ms | 0 - 17 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
44
+ | LiteHRNet | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 6.535 ms | 0 - 63 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
45
+ | LiteHRNet | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 4.496 ms | 0 - 17 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
46
+ | LiteHRNet | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 5.427 ms | 0 - 60 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
47
+ | LiteHRNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 2.824 ms | 0 - 68 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
48
+ | LiteHRNet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 3.398 ms | 0 - 55 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
49
+ | LiteHRNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 2.361 ms | 0 - 62 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
50
+ | LiteHRNet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 2.921 ms | 0 - 52 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
51
+ | LiteHRNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 2.138 ms | 0 - 61 MB | NPU | [LiteHRNet.tflite](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.tflite) |
52
+ | LiteHRNet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 2.771 ms | 1 - 56 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
53
  | LiteHRNet | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 5.849 ms | 4 - 4 MB | NPU | [LiteHRNet.onnx.zip](https://huggingface.co/qualcomm/LiteHRNet/blob/main/LiteHRNet.onnx.zip) |
54
 
55
 
 
66
  ```
67
 
68
 
69
+ ## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
70
 
71
+ Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.com/) with your
72
  Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
73
 
74
  With this API token, you can configure your client to run models on the cloud
 
76
  ```bash
77
  qai-hub configure --api_token API_TOKEN
78
  ```
79
+ Navigate to [docs](https://workbench.aihub.qualcomm.com/docs/) for more information.
80
 
81
 
82
 
 
187
  spot check the output with expected output.
188
 
189
  **Note**: This on-device profiling and inference requires access to Qualcomm®
190
+ AI Hub Workbench. [Sign up for access](https://myaccount.qualcomm.com/signup).
191
 
192
 
193