YAML Metadata Warning:The pipeline tag "gaze-estimation" is not in the official list: text-classification, token-classification, table-question-answering, question-answering, zero-shot-classification, translation, summarization, feature-extraction, text-generation, fill-mask, sentence-similarity, text-to-speech, text-to-audio, automatic-speech-recognition, audio-to-audio, audio-classification, audio-text-to-text, voice-activity-detection, depth-estimation, image-classification, object-detection, image-segmentation, text-to-image, image-to-text, image-to-image, image-to-video, unconditional-image-generation, video-classification, reinforcement-learning, robotics, tabular-classification, tabular-regression, tabular-to-text, table-to-text, multiple-choice, text-ranking, text-retrieval, time-series-forecasting, text-to-video, image-text-to-text, image-text-to-image, image-text-to-video, visual-question-answering, document-question-answering, zero-shot-image-classification, graph-ml, mask-generation, zero-shot-object-detection, text-to-3d, image-to-3d, image-feature-extraction, video-text-to-text, keypoint-detection, visual-document-retrieval, any-to-any, video-to-video, other
EyeGaze: Optimized for Qualcomm Devices
Predicts gaze direction (pitch, yaw) from 96x160 grayscale eye images using the EyeNet model.
This is based on the implementation of EyeGaze found here. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.
Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up to run these models on a hosted Qualcomm® device.
Getting Started
There are two ways to deploy this model on your device:
Option 1: Download Pre-Exported Models
Below are pre-exported model assets ready for deployment.
| Runtime | Precision | Chipset | SDK Versions | Download |
|---|---|---|---|---|
| ONNX | float | Universal | ONNX Runtime 1.25.0 | Download |
| ONNX | w8a16 | Universal | ONNX Runtime 1.25.0 | Download |
| QNN_DLC | float | Universal | QAIRT 2.45 | Download |
| TFLITE | float | Universal | Download |
For more device-specific assets and performance metrics, visit EyeGaze on Qualcomm® AI Hub.
Option 2: Export with Custom Configurations
Use the Qualcomm® AI Hub Models Python library to compile and export the model with your own:
- Custom weights (e.g., fine-tuned checkpoints)
- Custom input shapes
- Target device and runtime configurations
This option is ideal if you need to customize the model beyond the default configuration provided here.
See our repository for EyeGaze on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.gaze_estimation
Model Stats:
- Model checkpoint: checkpoint.pt
- Input resolution: 96x160
- Number of parameters: 2.58M
- Model size (float): 9.6MB
- Model size (w8a16): 3.3 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| EyeGaze | ONNX | float | Snapdragon® X2 Elite | 9.429 ms | 35 - 35 MB | CPU |
| EyeGaze | ONNX | float | Snapdragon® X Elite | 34.689 ms | 34 - 34 MB | CPU |
| EyeGaze | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 22.483 ms | 31 - 41 MB | CPU |
| EyeGaze | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 35.285 ms | 32 - 47 MB | CPU |
| EyeGaze | ONNX | float | Qualcomm® QCS8550 (Proxy) | 25.794 ms | 27 - 35 MB | CPU |
| EyeGaze | ONNX | float | Qualcomm® QCS8450 | 35.285 ms | 32 - 47 MB | CPU |
| EyeGaze | ONNX | float | Snapdragon® 8 Elite Mobile | 20.13 ms | 33 - 42 MB | CPU |
| EyeGaze | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 18.837 ms | 32 - 40 MB | CPU |
| EyeGaze | ONNX | float | Qualcomm® QCS9075 | 20.959 ms | 32 - 42 MB | CPU |
| EyeGaze | ONNX | float | Qualcomm® QCS8750 | 20.13 ms | 33 - 42 MB | CPU |
| EyeGaze | ONNX | float | Qualcomm® QCS7181 | 34.689 ms | 34 - 34 MB | CPU |
| EyeGaze | ONNX | w8a16 | Snapdragon® X2 Elite | 17.112 ms | 69 - 69 MB | CPU |
| EyeGaze | ONNX | w8a16 | Snapdragon® X Elite | 103.196 ms | 68 - 68 MB | CPU |
| EyeGaze | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 31.714 ms | 64 - 76 MB | CPU |
| EyeGaze | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 45.46 ms | 60 - 79 MB | CPU |
| EyeGaze | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 40.254 ms | 60 - 64 MB | CPU |
| EyeGaze | ONNX | w8a16 | Qualcomm® QCS8450 | 45.46 ms | 60 - 79 MB | CPU |
| EyeGaze | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 25.551 ms | 61 - 74 MB | CPU |
| EyeGaze | ONNX | w8a16 | Qualcomm® QCS9075 | 41.028 ms | 61 - 67 MB | CPU |
| EyeGaze | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 27.884 ms | 60 - 73 MB | CPU |
| EyeGaze | ONNX | w8a16 | Qualcomm® QCS8750 | 27.884 ms | 60 - 73 MB | CPU |
| EyeGaze | ONNX | w8a16 | Qualcomm® QCS7181 | 103.196 ms | 68 - 68 MB | CPU |
| EyeGaze | QNN_DLC | float | Snapdragon® X2 Elite | 9.645 ms | 11 - 11 MB | CPU |
| EyeGaze | QNN_DLC | float | Snapdragon® X Elite | 133.191 ms | 11 - 11 MB | CPU |
| EyeGaze | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 17.518 ms | 10 - 23 MB | CPU |
| EyeGaze | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 31.185 ms | 12 - 27 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® QCS8275 | 220.69 ms | 9 - 24 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 19.966 ms | 10 - 16 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® SA8775P | 29.143 ms | 10 - 21 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® SA8650P | 29.143 ms | 10 - 21 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® SA8255P | 29.143 ms | 10 - 21 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® QCS8450 | 31.185 ms | 12 - 27 MB | CPU |
| EyeGaze | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 14.334 ms | 10 - 26 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® SA8295P | 22.909 ms | 10 - 20 MB | CPU |
| EyeGaze | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 12.238 ms | 10 - 22 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® SA7255P | 220.69 ms | 9 - 24 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® QCS9075 | 66.986 ms | 68 - 142 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® QCS8750 | 14.334 ms | 10 - 26 MB | CPU |
| EyeGaze | QNN_DLC | float | Qualcomm® QCS7181 | 133.191 ms | 11 - 11 MB | CPU |
| EyeGaze | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 17.083 ms | 2 - 12 MB | CPU |
| EyeGaze | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 26.199 ms | 3 - 20 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® QCS8275 | 30.566 ms | 2 - 10 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 18.527 ms | 2 - 51 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® SA8775P | 25.302 ms | 2 - 7 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® SA8650P | 25.302 ms | 2 - 7 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® SA8255P | 25.302 ms | 2 - 7 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® QCS8450 | 26.199 ms | 3 - 20 MB | CPU |
| EyeGaze | TFLITE | float | Snapdragon® 8 Elite Mobile | 11.867 ms | 2 - 14 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® SA8295P | 16.696 ms | 2 - 7 MB | CPU |
| EyeGaze | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 6.287 ms | 2 - 10 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® SA7255P | 30.566 ms | 2 - 10 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® QCS9075 | 25.249 ms | 2 - 43 MB | CPU |
| EyeGaze | TFLITE | float | Qualcomm® QCS8750 | 11.867 ms | 2 - 14 MB | CPU |
License
- The license for the original implementation of EyeGaze can be found here.
References
Community
- Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
- For questions or feedback please reach out to us.