Name: headpose-estimator-apache-mxnet
Owner: AWS Samples
Description: Head Pose estimator using Apache MXNet. HeadPose_ResNet50_Tutorial.ipynb helps you to walk through an entire work flow of developing a CNN model from the scratch including data augmentation, fine-tuning, training, saving check-point model artifacts, validation and inference.
Created: 2018-02-06 19:36:56.0
Updated: 2018-02-13 19:08:20.0
Pushed: 2018-02-16 07:13:06.0
Homepage: null
Size: 1702
Language: Jupyter Notebook
GitHub Committers
| User | Most Recent Commit | # Commits |
|---|
Other Committers
| User | Most Recent Commit | # Commits |
|---|
Head Pose estimator using Apache MXNet. HeadPose_ResNet50_Tutorial.ipynb helps you to walk through an end-to-end work flow of developing a CNN model from the scratch including data augmentation, fine-tuning, saving check-point model artifacts, validation and inference.
Please run the following command first to prepare the input data file.
python2 preprocessingDataset_py2.py –num-data-aug 15 –aspect-ratio 1
Jupyter notebook to develop Headpose Estimator CNN model using Apache MXNet.
A set of SageMaker notebook and entry point script to develop the Headpose Estimator model on Amazon SageMaker.
HeadPose_SageMaker_PySDK.ipynb: SageMaker notebook to invoke an entry point python script.
EntryPt-headpose.py: An entry point python script to train Headpose Estimator model. This entry point script is analogous to HeadPose_ResNet50_Tutorial.ipynb.
EntryPt-headpose-wo-cv2.py: The entry point script without cv2.
Sample head images for inference test.
This library is licensed under the Apache 2.0 License.