Hugging Face's logo

qualcomm
/
EyeGaze

PyTorch
bu_auto
real_time
android
gaze-estimation
Model card Files
xet
 Community

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 QAIRT 2.45, ONNX Runtime 1.27.1 Download
ONNX w8a16 Universal QAIRT 2.45, ONNX Runtime 1.27.1 Download
QNN_DLC float Universal QAIRT 2.45 Download
QNN_DLC w8a16 Universal QAIRT 2.45 Download
TFLITE float Universal QAIRT 2.45 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 0.618 ms 8 - 8 MB NPU
EyeGaze ONNX float Snapdragon® X Elite 1.266 ms 8 - 8 MB NPU
EyeGaze ONNX float Snapdragon® 8 Gen 3 Mobile 0.87 ms 0 - 60 MB NPU
EyeGaze ONNX float Snapdragon® 8 Gen 1 Mobile 1.721 ms 0 - 65 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.25 ms 0 - 82 MB NPU
EyeGaze ONNX float Qualcomm® QCS8450 1.721 ms 0 - 65 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ IQ-9075 1.71 ms 0 - 3 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ IQ-X7181 1.266 ms 8 - 8 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ Q-8750 0.688 ms 0 - 51 MB NPU
EyeGaze ONNX float Snapdragon® 8 Elite Mobile 0.688 ms 0 - 51 MB NPU
EyeGaze ONNX float Snapdragon® 8 Elite Gen 5 Mobile 0.596 ms 0 - 52 MB NPU
EyeGaze ONNX w8a16 Snapdragon® X2 Elite 0.503 ms 3 - 3 MB NPU
EyeGaze ONNX w8a16 Snapdragon® X Elite 0.942 ms 0 - 0 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 0.622 ms 0 - 83 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 8 Gen 1 Mobile 1.115 ms 1 - 81 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ QCS6490 4.216 ms 2 - 4 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.922 ms 0 - 6 MB NPU
EyeGaze ONNX w8a16 Qualcomm® QCS8450 1.115 ms 1 - 81 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 1.063 ms 0 - 3 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ IQ-X7181 0.942 ms 0 - 0 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ Q-8750 0.489 ms 0 - 68 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 8 Elite Mobile 0.489 ms 0 - 68 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 8 Elite Gen 5 Mobile 0.456 ms 0 - 65 MB NPU
EyeGaze QNN_DLC float Snapdragon® X2 Elite 0.74 ms 0 - 0 MB NPU
EyeGaze QNN_DLC float Snapdragon® X Elite 1.2 ms 0 - 0 MB NPU
EyeGaze QNN_DLC float Snapdragon® 8 Gen 3 Mobile 0.744 ms 0 - 52 MB NPU
EyeGaze QNN_DLC float Snapdragon® 8 Gen 1 Mobile 1.498 ms 0 - 55 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ QCS8275 4.171 ms 0 - 40 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.007 ms 0 - 2 MB NPU
EyeGaze QNN_DLC float Qualcomm® SA8775P 1.638 ms 0 - 41 MB NPU
EyeGaze QNN_DLC float Qualcomm® SA8650P 1.638 ms 0 - 41 MB NPU
EyeGaze QNN_DLC float Qualcomm® SA8255P 1.638 ms 0 - 41 MB NPU
EyeGaze QNN_DLC float Qualcomm® QCS8450 1.498 ms 0 - 55 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 1.49 ms 0 - 3 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 1.2 ms 0 - 0 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ Q-8750 0.593 ms 0 - 37 MB NPU
EyeGaze QNN_DLC float Qualcomm® SA7255P 4.171 ms 0 - 40 MB NPU
EyeGaze QNN_DLC float Qualcomm® SA8295P 1.825 ms 0 - 38 MB NPU
EyeGaze QNN_DLC float Snapdragon® 8 Elite Mobile 0.593 ms 0 - 37 MB NPU
EyeGaze QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 0.547 ms 0 - 39 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® X2 Elite 0.788 ms 0 - 0 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® X Elite 1.436 ms 0 - 0 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® 8 Gen 3 Mobile 0.861 ms 0 - 59 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® 8 Gen 1 Mobile 1.853 ms 0 - 62 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS8275 2.783 ms 0 - 46 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.222 ms 0 - 1 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® SA8775P 1.526 ms 0 - 48 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® SA8650P 1.526 ms 0 - 48 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® SA8255P 1.526 ms 0 - 48 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® QCS8450 1.853 ms 0 - 62 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-9075 1.387 ms 2 - 4 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-X7181 1.436 ms 0 - 0 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-8750 0.69 ms 0 - 50 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® SA7255P 2.783 ms 0 - 46 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® SA8295P 2.286 ms 0 - 44 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® 8 Elite Mobile 0.69 ms 0 - 50 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® 8 Elite Gen 5 Mobile 0.524 ms 0 - 44 MB NPU
EyeGaze TFLITE float Snapdragon® 8 Gen 3 Mobile 1.27 ms 0 - 170 MB NPU
EyeGaze TFLITE float Snapdragon® 8 Gen 1 Mobile 2.924 ms 2 - 175 MB NPU
EyeGaze TFLITE float Qualcomm® Dragonwing™ QCS8275 5.298 ms 2 - 159 MB NPU
EyeGaze TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.808 ms 0 - 52 MB NPU
EyeGaze TFLITE float Qualcomm® SA8775P 2.478 ms 0 - 160 MB NPU
EyeGaze TFLITE float Qualcomm® SA8650P 2.478 ms 0 - 160 MB NPU
EyeGaze TFLITE float Qualcomm® SA8255P 2.478 ms 0 - 160 MB NPU
EyeGaze TFLITE float Qualcomm® QCS8450 2.924 ms 2 - 175 MB NPU
EyeGaze TFLITE float Qualcomm® Dragonwing™ IQ-9075 2.284 ms 0 - 11 MB NPU
EyeGaze TFLITE float Qualcomm® Dragonwing™ Q-8750 1.114 ms 0 - 156 MB NPU
EyeGaze TFLITE float Qualcomm® SA7255P 5.298 ms 2 - 159 MB NPU
EyeGaze TFLITE float Qualcomm® SA8295P 2.879 ms 0 - 156 MB NPU
EyeGaze TFLITE float Snapdragon® 8 Elite Mobile 1.114 ms 0 - 156 MB NPU
EyeGaze TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 0.944 ms 0 - 159 MB NPU

License

  • The license for the original implementation of EyeGaze can be found here.

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support