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 2.50, ONNX Runtime 1.27.1 Download
ONNX w8a16 Universal QAIRT 2.50, ONNX Runtime 1.27.1 Download
QNN_DLC float Universal QAIRT 2.50 Download
QNN_DLC w8a16 Universal QAIRT 2.50 Download
TFLITE float Universal QAIRT 2.50 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:

  • Input resolution: 96x160
  • Model checkpoint: checkpoint.pt
  • Model size (float): 9.6MB
  • Model size (w8a16): 3.3 MB
  • Number of parameters: 2.58M

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
EyeGaze ONNX float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.588 ms 0 - 53 MB NPU
EyeGaze ONNX float Snapdragon® 8 Elite For Galaxy Mobile 0.662 ms 0 - 58 MB NPU
EyeGaze ONNX float Snapdragon® X2 Elite 0.626 ms 8 - 8 MB NPU
EyeGaze ONNX float Snapdragon® X Elite 1.243 ms 8 - 8 MB NPU
EyeGaze ONNX float Snapdragon® 8 Gen 3 Mobile 0.844 ms 0 - 68 MB NPU
EyeGaze ONNX float Snapdragon® 8 Gen 1 Mobile 1.673 ms 2 - 71 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ IQ-8275 1.658 ms 0 - 4 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.211 ms 0 - 8 MB NPU
EyeGaze ONNX float Qualcomm® QCS8450 1.673 ms 2 - 71 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ IQ-9075 1.646 ms 0 - 3 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ IQ-X7181 1.243 ms 8 - 8 MB NPU
EyeGaze ONNX float Qualcomm® Dragonwing™ Q-8750 0.662 ms 0 - 58 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.478 ms 0 - 71 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 8 Elite For Galaxy Mobile 0.491 ms 0 - 71 MB NPU
EyeGaze ONNX w8a16 Snapdragon® X2 Elite 0.529 ms 3 - 3 MB NPU
EyeGaze ONNX w8a16 Snapdragon® X Elite 0.945 ms 0 - 0 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 0.623 ms 0 - 81 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 8 Gen 1 Mobile 1.092 ms 1 - 83 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ QCS6490 3.003 ms 0 - 3 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ IQ-8275 1.013 ms 0 - 4 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.929 ms 0 - 63 MB NPU
EyeGaze ONNX w8a16 Qualcomm® QCS8450 1.092 ms 1 - 83 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 1.068 ms 0 - 3 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ IQ-X7181 0.945 ms 0 - 0 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ Q-6690 5.946 ms 2 - 178 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ Q-7790 1.105 ms 0 - 175 MB NPU
EyeGaze ONNX w8a16 Qualcomm® Dragonwing™ Q-8750 0.491 ms 0 - 71 MB NPU
EyeGaze ONNX w8a16 Snapdragon® 7 Gen 4 Mobile 1.105 ms 0 - 175 MB NPU
EyeGaze QNN_DLC float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.718 ms 0 - 44 MB NPU
EyeGaze QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 0.807 ms 0 - 43 MB NPU
EyeGaze QNN_DLC float Snapdragon® X2 Elite 0.965 ms 0 - 0 MB NPU
EyeGaze QNN_DLC float Snapdragon® X Elite 1.721 ms 0 - 0 MB NPU
EyeGaze QNN_DLC float Snapdragon® 8 Gen 3 Mobile 1.03 ms 0 - 58 MB NPU
EyeGaze QNN_DLC float Snapdragon® 8 Gen 1 Mobile 1.952 ms 0 - 62 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ IQ-8275 1.882 ms 0 - 4 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.5 ms 0 - 2 MB NPU
EyeGaze QNN_DLC float Qualcomm® QCS8450 1.952 ms 0 - 62 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 1.955 ms 0 - 4 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 1.721 ms 0 - 0 MB NPU
EyeGaze QNN_DLC float Qualcomm® Dragonwing™ Q-8750 0.807 ms 0 - 43 MB NPU
EyeGaze QNN_DLC float Qualcomm® SA8295P 2.299 ms 0 - 43 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.66 ms 0 - 51 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® 8 Elite For Galaxy Mobile 0.79 ms 0 - 49 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® X2 Elite 0.944 ms 0 - 0 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® X Elite 1.747 ms 0 - 0 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® 8 Gen 3 Mobile 1.04 ms 0 - 63 MB NPU
EyeGaze QNN_DLC w8a16 Snapdragon® 8 Gen 1 Mobile 2.172 ms 0 - 67 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-8275 1.463 ms 0 - 4 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.52 ms 0 - 2 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® SA8650P 1.532 ms 0 - 48 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® SA8255P 1.532 ms 0 - 48 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® QCS8450 2.172 ms 0 - 67 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-9075 1.67 ms 0 - 3 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-X7181 1.747 ms 0 - 0 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-8750 0.79 ms 0 - 49 MB NPU
EyeGaze QNN_DLC w8a16 Qualcomm® SA8295P 2.575 ms 0 - 49 MB NPU
EyeGaze TFLITE float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.666 ms 0 - 42 MB NPU
EyeGaze TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 0.766 ms 0 - 48 MB NPU
EyeGaze TFLITE float Snapdragon® 8 Gen 3 Mobile 0.95 ms 0 - 56 MB NPU
EyeGaze TFLITE float Snapdragon® 8 Gen 1 Mobile 1.838 ms 0 - 63 MB NPU
EyeGaze TFLITE float Qualcomm® Dragonwing™ IQ-8275 1.779 ms 0 - 10 MB NPU
EyeGaze TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.383 ms 0 - 3 MB NPU
EyeGaze TFLITE float Qualcomm® SA8775P 12.676 ms 0 - 31 MB GPU
EyeGaze TFLITE float Qualcomm® SA8650P 12.676 ms 0 - 31 MB GPU
EyeGaze TFLITE float Qualcomm® SA8255P 12.676 ms 0 - 31 MB GPU
EyeGaze TFLITE float Qualcomm® QCS8450 1.838 ms 0 - 63 MB NPU
EyeGaze TFLITE float Qualcomm® Dragonwing™ IQ-9075 1.818 ms 0 - 10 MB NPU
EyeGaze TFLITE float Qualcomm® Dragonwing™ Q-8750 0.766 ms 0 - 48 MB NPU
EyeGaze TFLITE float Qualcomm® SA7255P 22.469 ms 0 - 30 MB GPU
EyeGaze TFLITE float Qualcomm® SA8295P 2.099 ms 0 - 42 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