Hugging Face's logo

qualcomm
/
EyeGaze

PyTorch
bu_auto
real_time
android
gaze-estimation
Model card Files
xet
 Community
qaihm-bot commited on
Commit
ba5b558
·
verified ·
1 Parent(s): 8e75d0c

v0.48.0

Browse files

See https://github.com/qualcomm/ai-hub-models/releases/v0.48.0 for changelog.

Files changed (1) hide show
  1. README.md +36 -36
README.md CHANGED
@@ -16,7 +16,7 @@ pipeline_tag: gaze-estimation
16
  Predicts gaze direction (pitch, yaw) from 96x160 grayscale eye images using the EyeNet model.
17
 
18
  This is based on the implementation of EyeGaze found [here](https://github.com/david-wb/gaze-estimation).
19
- This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/eyegaze) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
20
 
21
  Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) to run these models on a hosted Qualcomm® device.
22
 
@@ -29,24 +29,24 @@ Below are pre-exported model assets ready for deployment.
29
 
30
  | Runtime | Precision | Chipset | SDK Versions | Download |
31
  |---|---|---|---|---|
32
- | ONNX | float | Universal | ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/eyegaze/releases/v0.47.0/eyegaze-onnx-float.zip)
33
- | ONNX | w8a16 | Universal | ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/eyegaze/releases/v0.47.0/eyegaze-onnx-w8a16.zip)
34
- | QNN_DLC | float | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/eyegaze/releases/v0.47.0/eyegaze-qnn_dlc-float.zip)
35
- | TFLITE | float | Universal | TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/eyegaze/releases/v0.47.0/eyegaze-tflite-float.zip)
36
 
37
  For more device-specific assets and performance metrics, visit **[EyeGaze on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/eyegaze)**.
38
 
39
 
40
  ### Option 2: Export with Custom Configurations
41
 
42
- Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/eyegaze) Python library to compile and export the model with your own:
43
  - Custom weights (e.g., fine-tuned checkpoints)
44
  - Custom input shapes
45
  - Target device and runtime configurations
46
 
47
  This option is ideal if you need to customize the model beyond the default configuration provided here.
48
 
49
- See our repository for [EyeGaze on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/eyegaze) for usage instructions.
50
 
51
  ## Model Details
52
 
@@ -62,35 +62,35 @@ See our repository for [EyeGaze on GitHub](https://github.com/quic/ai-hub-models
62
  ## Performance Summary
63
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
64
  |---|---|---|---|---|---|---
65
- | EyeGaze | ONNX | float | Snapdragon® X Elite | 10.083 ms | 34 - 34 MB | CPU
66
- | EyeGaze | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 21.665 ms | 32 - 42 MB | CPU
67
- | EyeGaze | ONNX | float | Qualcomm® QCS8550 (Proxy) | 25.229 ms | 27 - 40 MB | CPU
68
- | EyeGaze | ONNX | float | Qualcomm® QCS9075 | 21.013 ms | 32 - 42 MB | CPU
69
- | EyeGaze | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 20.067 ms | 34 - 46 MB | CPU
70
- | EyeGaze | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 18.835 ms | 32 - 42 MB | CPU
71
- | EyeGaze | ONNX | float | Snapdragon® X2 Elite | 9.741 ms | 34 - 34 MB | CPU
72
- | EyeGaze | ONNX | w8a16 | Snapdragon® X Elite | 17.509 ms | 101 - 101 MB | CPU
73
- | EyeGaze | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 31.52 ms | 70 - 85 MB | CPU
74
- | EyeGaze | ONNX | w8a16 | Qualcomm® QCS6490 | 140.663 ms | 68 - 73 MB | CPU
75
- | EyeGaze | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 37.518 ms | 69 - 72 MB | CPU
76
- | EyeGaze | ONNX | w8a16 | Qualcomm® QCS9075 | 40.83 ms | 71 - 76 MB | CPU
77
- | EyeGaze | ONNX | w8a16 | Qualcomm® QCM6690 | 68.264 ms | 71 - 82 MB | CPU
78
- | EyeGaze | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 27.303 ms | 70 - 85 MB | CPU
79
- | EyeGaze | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 60.626 ms | 59 - 70 MB | CPU
80
- | EyeGaze | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 26.07 ms | 69 - 84 MB | CPU
81
- | EyeGaze | ONNX | w8a16 | Snapdragon® X2 Elite | 17.889 ms | 102 - 102 MB | CPU
82
- | EyeGaze | QNN_DLC | float | Snapdragon® X Elite | 35.483 ms | 13 - 13 MB | CPU
83
- | EyeGaze | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 35.931 ms | 11 - 23 MB | CPU
84
- | EyeGaze | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 232.353 ms | 22 - 34 MB | CPU
85
- | EyeGaze | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 44.999 ms | 35 - 41 MB | CPU
86
- | EyeGaze | QNN_DLC | float | Qualcomm® SA8775P | 57.217 ms | 36 - 49 MB | CPU
87
- | EyeGaze | QNN_DLC | float | Qualcomm® QCS9075 | 57.227 ms | 64 - 137 MB | CPU
88
- | EyeGaze | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 66.634 ms | 11 - 23 MB | CPU
89
- | EyeGaze | QNN_DLC | float | Qualcomm® SA7255P | 232.353 ms | 22 - 34 MB | CPU
90
- | EyeGaze | QNN_DLC | float | Qualcomm® SA8295P | 43.601 ms | 33 - 42 MB | CPU
91
- | EyeGaze | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 39.436 ms | 37 - 50 MB | CPU
92
- | EyeGaze | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 24.269 ms | 11 - 24 MB | CPU
93
- | EyeGaze | QNN_DLC | float | Snapdragon® X2 Elite | 41.795 ms | 14 - 14 MB | CPU
94
 
95
  ## License
96
  * The license for the original implementation of EyeGaze can be found
 
16
  Predicts gaze direction (pitch, yaw) from 96x160 grayscale eye images using the EyeNet model.
17
 
18
  This is based on the implementation of EyeGaze found [here](https://github.com/david-wb/gaze-estimation).
19
+ This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/eyegaze) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
20
 
21
  Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) to run these models on a hosted Qualcomm® device.
22
 
 
29
 
30
  | Runtime | Precision | Chipset | SDK Versions | Download |
31
  |---|---|---|---|---|
32
+ | ONNX | float | Universal | ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/eyegaze/releases/v0.48.0/eyegaze-onnx-float.zip)
33
+ | ONNX | w8a16 | Universal | ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/eyegaze/releases/v0.48.0/eyegaze-onnx-w8a16.zip)
34
+ | QNN_DLC | float | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/eyegaze/releases/v0.48.0/eyegaze-qnn_dlc-float.zip)
35
+ | TFLITE | float | Universal | TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/eyegaze/releases/v0.48.0/eyegaze-tflite-float.zip)
36
 
37
  For more device-specific assets and performance metrics, visit **[EyeGaze on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/eyegaze)**.
38
 
39
 
40
  ### Option 2: Export with Custom Configurations
41
 
42
+ Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/eyegaze) Python library to compile and export the model with your own:
43
  - Custom weights (e.g., fine-tuned checkpoints)
44
  - Custom input shapes
45
  - Target device and runtime configurations
46
 
47
  This option is ideal if you need to customize the model beyond the default configuration provided here.
48
 
49
+ See our repository for [EyeGaze on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/eyegaze) for usage instructions.
50
 
51
  ## Model Details
52
 
 
62
  ## Performance Summary
63
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
64
  |---|---|---|---|---|---|---
65
+ | EyeGaze | ONNX | float | Snapdragon® X2 Elite | 8.981 ms | 35 - 35 MB | CPU
66
+ | EyeGaze | ONNX | float | Snapdragon® X Elite | 9.533 ms | 34 - 34 MB | CPU
67
+ | EyeGaze | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 21.778 ms | 31 - 41 MB | CPU
68
+ | EyeGaze | ONNX | float | Qualcomm® QCS8550 (Proxy) | 31.181 ms | 32 - 46 MB | CPU
69
+ | EyeGaze | ONNX | float | Qualcomm® QCS9075 | 21.033 ms | 32 - 42 MB | CPU
70
+ | EyeGaze | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 19.985 ms | 32 - 40 MB | CPU
71
+ | EyeGaze | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 18.935 ms | 32 - 42 MB | CPU
72
+ | EyeGaze | ONNX | w8a16 | Snapdragon® X2 Elite | 17.195 ms | 102 - 102 MB | CPU
73
+ | EyeGaze | ONNX | w8a16 | Snapdragon® X Elite | 17.537 ms | 101 - 101 MB | CPU
74
+ | EyeGaze | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 31.803 ms | 69 - 83 MB | CPU
75
+ | EyeGaze | ONNX | w8a16 | Qualcomm® QCS6490 | 146.93 ms | 68 - 74 MB | CPU
76
+ | EyeGaze | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 38.244 ms | 70 - 74 MB | CPU
77
+ | EyeGaze | ONNX | w8a16 | Qualcomm® QCS9075 | 40.879 ms | 68 - 74 MB | CPU
78
+ | EyeGaze | ONNX | w8a16 | Qualcomm® QCM6690 | 68.17 ms | 69 - 80 MB | CPU
79
+ | EyeGaze | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 27.824 ms | 71 - 84 MB | CPU
80
+ | EyeGaze | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 60.413 ms | 58 - 69 MB | CPU
81
+ | EyeGaze | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 25.889 ms | 69 - 84 MB | CPU
82
+ | EyeGaze | QNN_DLC | float | Snapdragon® X2 Elite | 40.969 ms | 14 - 14 MB | CPU
83
+ | EyeGaze | QNN_DLC | float | Snapdragon® X Elite | 28.924 ms | 13 - 13 MB | CPU
84
+ | EyeGaze | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 35.31 ms | 14 - 26 MB | CPU
85
+ | EyeGaze | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 238.831 ms | 22 - 35 MB | CPU
86
+ | EyeGaze | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 46.681 ms | 36 - 41 MB | CPU
87
+ | EyeGaze | QNN_DLC | float | Qualcomm® SA8775P | 51.057 ms | 36 - 45 MB | CPU
88
+ | EyeGaze | QNN_DLC | float | Qualcomm® QCS9075 | 56.177 ms | 76 - 149 MB | CPU
89
+ | EyeGaze | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 61.833 ms | 11 - 25 MB | CPU
90
+ | EyeGaze | QNN_DLC | float | Qualcomm® SA7255P | 238.831 ms | 22 - 35 MB | CPU
91
+ | EyeGaze | QNN_DLC | float | Qualcomm® SA8295P | 42.662 ms | 34 - 42 MB | CPU
92
+ | EyeGaze | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 38.838 ms | 33 - 43 MB | CPU
93
+ | EyeGaze | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 26.141 ms | 37 - 51 MB | CPU
94
 
95
  ## License
96
  * The license for the original implementation of EyeGaze can be found