Hugging Face's logo

qualcomm
/
RTMDet

Object Detection
PyTorch
real_time
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
535fe73
·
verified ·
1 Parent(s): c4ff89e

v0.42.0

Browse files

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

Files changed (1) hide show
  1. README.md +21 -21
README.md CHANGED
@@ -36,23 +36,23 @@ More details on model performance across various devices, can be found
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
- | RTMDet | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 85.367 ms | 0 - 66 MB | NPU | -- |
40
- | RTMDet | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 32.406 ms | 0 - 112 MB | NPU | -- |
41
- | RTMDet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 15.905 ms | 0 - 17 MB | NPU | -- |
42
- | RTMDet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 14.328 ms | 0 - 121 MB | NPU | -- |
43
- | RTMDet | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 23.93 ms | 0 - 66 MB | NPU | -- |
44
- | RTMDet | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 85.367 ms | 0 - 66 MB | NPU | -- |
45
- | RTMDet | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 16.982 ms | 0 - 13 MB | NPU | -- |
46
- | RTMDet | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 35.479 ms | 0 - 74 MB | NPU | -- |
47
- | RTMDet | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 16.623 ms | 0 - 13 MB | NPU | -- |
48
- | RTMDet | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 23.93 ms | 0 - 66 MB | NPU | -- |
49
- | RTMDet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 12.201 ms | 0 - 107 MB | NPU | -- |
50
- | RTMDet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 10.617 ms | 4 - 48 MB | NPU | -- |
51
- | RTMDet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 9.41 ms | 0 - 71 MB | NPU | -- |
52
- | RTMDet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 8.329 ms | 2 - 45 MB | NPU | -- |
53
- | RTMDet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 7.218 ms | 0 - 73 MB | NPU | -- |
54
- | RTMDet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 6.242 ms | 4 - 47 MB | NPU | -- |
55
- | RTMDet | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 14.52 ms | 51 - 51 MB | NPU | -- |
56
 
57
 
58
 
@@ -67,9 +67,9 @@ pip install "qai-hub-models[rtmdet]"
67
  ```
68
 
69
 
70
- ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
71
 
72
- Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
73
  Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
74
 
75
  With this API token, you can configure your client to run models on the cloud
@@ -77,7 +77,7 @@ hosted devices.
77
  ```bash
78
  qai-hub configure --api_token API_TOKEN
79
  ```
80
- Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information.
81
 
82
 
83
 
@@ -188,7 +188,7 @@ With the output of the model, you can compute like PSNR, relative errors or
188
  spot check the output with expected output.
189
 
190
  **Note**: This on-device profiling and inference requires access to Qualcomm®
191
- AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
192
 
193
 
194
 
 
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
+ | RTMDet | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 85.262 ms | 0 - 68 MB | NPU | -- |
40
+ | RTMDet | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 32.666 ms | 0 - 113 MB | NPU | -- |
41
+ | RTMDet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 15.88 ms | 0 - 14 MB | NPU | -- |
42
+ | RTMDet | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 14.188 ms | 5 - 21 MB | NPU | -- |
43
+ | RTMDet | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 108.091 ms | 0 - 68 MB | NPU | -- |
44
+ | RTMDet | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 85.262 ms | 0 - 68 MB | NPU | -- |
45
+ | RTMDet | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 16.255 ms | 0 - 13 MB | NPU | -- |
46
+ | RTMDet | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 35.567 ms | 0 - 76 MB | NPU | -- |
47
+ | RTMDet | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 15.982 ms | 0 - 17 MB | NPU | -- |
48
+ | RTMDet | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 108.091 ms | 0 - 68 MB | NPU | -- |
49
+ | RTMDet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 12.22 ms | 0 - 109 MB | NPU | -- |
50
+ | RTMDet | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 10.633 ms | 5 - 52 MB | NPU | -- |
51
+ | RTMDet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 9.39 ms | 0 - 72 MB | NPU | -- |
52
+ | RTMDet | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 8.324 ms | 3 - 45 MB | NPU | -- |
53
+ | RTMDet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 7.195 ms | 0 - 74 MB | NPU | -- |
54
+ | RTMDet | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 6.248 ms | 4 - 49 MB | NPU | -- |
55
+ | RTMDet | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 14.486 ms | 51 - 51 MB | NPU | -- |
56
 
57
 
58
 
 
67
  ```
68
 
69
 
70
+ ## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
71
 
72
+ Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.com/) with your
73
  Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
74
 
75
  With this API token, you can configure your client to run models on the cloud
 
77
  ```bash
78
  qai-hub configure --api_token API_TOKEN
79
  ```
80
+ Navigate to [docs](https://workbench.aihub.qualcomm.com/docs/) for more information.
81
 
82
 
83
 
 
188
  spot check the output with expected output.
189
 
190
  **Note**: This on-device profiling and inference requires access to Qualcomm®
191
+ AI Hub Workbench. [Sign up for access](https://myaccount.qualcomm.com/signup).
192
 
193
 
194