3.2: Introduction to TensorFlow and Keras with Anaconda Python (Module 3, Part 2)

Overview of deep learning tools. How to construct a neural network with TensorFlow and Keras for simple classification and regression. This video is part of a course that is taught in a hybrid format at Washington University in St. Louis; however, all the information is online and you can easily follow along. T81-558: Application of Deep Learning, at Washington University in St. Louis

Please subscribe and comment!

Follow me:

More links:

[zachdjarchitect]

great video Jeff! Really like seeing the predictive outputs which you don't see in other Tensorflow/Keras tutorials

[taliehtabatabaei]

Great course, thanks!
I am having problems with tensorflow. I followed your instruction for installing tensorflow version 1.12 . I believe it is properly installed since "print(f"Tensor Flow Version: {tf.__version__}")" gives the right output and all the helpful functions and other libraries work well. However, tensorflow functions don't work. And I noticed in your notes that now you have Tensorflow version 2.0 installed. Could that be a problem or it should work fine with 1.12 as well?
For example I get this error for the fractal codes: "ValueError: Cannot execute operation using `run()`: No default session is registered. Use `with sess.as_default():` or pass an explicit session to `run(session=sess)`"
Other ones don't work either. I would appreciate it if you help.

[luizcarlosmarquesjunior.]

Hello Professor Heaton, thanks for the classes I am having some issues with the Iris example. When I implement the code I receive the following message: NameError Traceback (most recent call last)
in <module>()
34 df=pd.read_csv(io.StringIO(requests.get(url).content.decode('utf-8')), na_values=['NA', '?'])
35
---> 36 species = encode_text_index(df, "species")
37 x, y = to_xy(df, "species")
38

in encode_text_index(df, name)
10 # Encode text values to indexes(i.e. [1], [2], [3] for red, green, blue).
11 def encode_text_index(df, name):
---> 12 le = preprocessing.LabelEncoder()
13 df[name] = le.fit_transform(df[name])
14 return le.classes_

NameError: name 'preprocessing' is not defined

could you help me please.

Best regards.

[symnshah]

Sir, I have a question about the Neural Network to learn about XOR. Why there is no activation (softmax) for the second Dense layer (model.add(Dense(1)) )?

[erickmiller]

Great course Jeff! I've noticed though, that the version of tensorflow that you are using in the fractal example is 1.3.0, while you're most recent software package update indicates installation of 1.12.0. I'm on the latter, and I couldn't get the Fractal example to run without a session error. In fact, there are a few others in the comments here who are having the same issue.


Just wondering if you're looking into this and if you came up with the fix for 1.12.0


Thanks!

[awaisasghar]

great video sir jeff!! i execute the code tensorflow graph on my jupyter notebook. it end up with error i.e.
Cannot execute operation using `run()`: No default session is registered. Use `with sess.as_default():` or pass an explicit session to `run(session=sess)`
when i pass the arg session=sess in run. agrin error
"Attempted to use a closed Session"
please help me solve my problem.