Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: This project aims to construct a predictive model using a TensorFlow convolutional neural network (CNN) and document the end-to-end steps using a template. The Yosemite Summer vs. Winter dataset is a binary classification situation where we attempt to predict one of the two possible outcomes.

INTRODUCTION: The CycleGAN dataset collection contains datasets that consist of images from two classes A and B (for example, apple vs. orange, horses vs. zebras, and so on). The researchers used the images to train machine learning models for research work in the area of General Adversarial Networks.

From iteration Take1, we constructed and tuned a machine learning model for this dataset using TensorFlow with a simple VGG-3 network. We also observed the best result that we could obtain using the test dataset. The final model from this iteration became our baseline model for future iterations.

From iteration Take2, we constructed and tuned a machine learning model for this dataset using TensorFlow with a VGG-16 network. We also observed the best result that we could obtain using the test dataset. The final model from this iteration became our baseline model for future iterations.

From iteration Take3, we constructed and tuned a machine learning model for this dataset using TensorFlow with a Inception V3 network. We also observed the best result that we could obtain using the test dataset. The final model from this iteration became our baseline model for future iterations.

In this Take4 iteration, we will construct and tune a machine learning model for this dataset using TensorFlow with a ResNet50 V2 network. We will also observe the best result that we can obtain using the test dataset and compare the performance with the baseline model.

ANALYSIS: From iteration Take1, the baseline model’s performance achieved an accuracy score of 99.00% after 20 epochs using the training dataset. The final model also processed the test dataset with 74.95% accuracy.

From iteration Take2, the VGG-16 model’s performance achieved an accuracy score of 74.44% after 20 epochs using the training dataset. The final model also processed the test dataset with 70.02% accuracy.

From iteration Take3, the Inception V3 model’s performance achieved an accuracy score of 83.36% after 20 epochs using the training dataset. The final model also processed the test dataset with 52.83% accuracy.

In this Take4 iteration, the ResNet50 V2 model’s performance achieved an accuracy score of 57.48% after 20 epochs using the training dataset. The final model also processed the test dataset with 46.80% accuracy.

CONCLUSION: In this iteration, the ResNet50 V2 CNN model did not appear to be suitable for modeling this dataset. We should consider experimenting with TensorFlow for further modeling.

Dataset Used: Yosemite Summer vs. Winter Dataset

Dataset ML Model: Binary image classification with numerical attributes

Dataset Reference: https://people.eecs.berkeley.edu/%7Etaesung_park/CycleGAN/datasets/

One potential source of performance benchmarks: https://arxiv.org/abs/1703.10593 or https://junyanz.github.io/CycleGAN/

The HTML formatted report can be found here on GitHub.