(Hands-On) Identify Ideal Model for Tuning
These results are achieved when running with the large datasets. If you are running the pipeline with the small datasets included in the Kale repo, the results will be less than ideal.
1. View the cnn-from-scratch logs
Take a look at the logs of the cnn-from-scratch step. This is the step where we trained a Convolutional Neural Network (CNN) from scratch. Notice that the trained model has a very low accuracy and this step took a long time to complete.
2. View the cnn_vgg16 logs
Take a look at the logs of the cnn_vgg16 step. In this step, we used transfer learning on the pre-trained VGG-16 model — a neural network trained by the Visual Geometry Group (VGG). The accuracy is much higher than the previous model, but we can still do better.
3. View the cnn_resnet50 logs
Now, take a look at the logs of the cnn-resnet50 step. In this step, we used transfer learning on the pre-trained ResNet-50 model. The accuracy is much higher. This is the model we should use for the rest of this course.
