Example questions/assignments for tensorflow certification exam

[JaapHaitsma]

@mrdbourke what I understood by searching on the web is that the certification exam consists of making 5 models that you need to upload and google will then grade them. Can you give some examples of how this exam works? I'm not asking you to violate your NDA but you could give an example based on one of the models you discuss during the course. E.g. you get this dataset and this notebook and now you have to make a model that has at least 90% accuracy

[mrdbourke]

Hey Jaap!

Your discoveries are correct (as long as the research your doing is talking about the same exam I took).

As for how they work.

You're given a dataset of some kind (these are relatively small) and then have to build a TensorFlow model to perform at some decent performance threshold (don't overthink this, if you think your model is doing pretty well given the data, you're probably right).

The questions will be similar to the style of script in: https://github.com/mrdbourke/tensorflow-deep-learning/blob/main/extras/image_classification_test.py

More specifically:

import tensorflow as tf
from tensorflow.keras import datasets, layers

# Check version of TensorFlow (exam requires a certain version)
# See for version: https://www.tensorflow.org/extras/cert/Setting_Up_TF_Developer_Certificate_Exam.pdf 
print(tf.__version__)

# Get data
(train_images, train_labels), (test_images, test_labels) = datasets.fashion_mnist.load_data()

# Normalize images (get values between 0 & 1)
train_images, test_images = train_images / 255.0, test_images / 255.0 

# Check shape of input data
# print(train_images.shape)
# print(train_labels.shape)

### YOUR CODE HERE ###

# Build a model which performs at ~90% accuracy on the test dataset

# Compile model 


# Fit model

# Evaluate model 

### END YOUR CODE ###

# Save model to current working directory
model.save("test_image_model.h5")

Notice the "YOUR CODE HERE" section.

For more details on how to prepare, I'd checkout the resources mentioned in the "Preparing for the TensorFlow Developer Certification" chapter of the course book: https://dev.mrdbourke.com/tensorflow-deep-learning/11_passing_the_tensorflow_developer_certification_exam/