You can run the validation routine with the following command:

python eval_image_classifier.py --alsologtostderr --checkpoint_path=D:datasetsdiabeticcheckpointsmodel.ckpt-92462 --dataset_name=diabetic --dataset_split_name=validation --dataset_dir=D:datasetsdiabetic	frecords --model_name=inception_v3

As you can see, the current accuracy is about 75 percent. In the Going further section, we will give you some ideas to improve this accuracy.

Now, we will look at the TensorBoard to visualize the training process.