possible issue at: tensorflow_v2/notebooks/3_NeuralNetworks/autoencoder.ipynb

[nucccc]

Referring to the notebook:

https://github.com/aymericdamien/TensorFlow-Examples/blob/master/tensorflow_v2/notebooks/3_NeuralNetworks/autoencoder.ipynb

In the following code portion (8th code block of the notebook):

`# Optimization process.
def run_optimization(x):
# Wrap computation inside a GradientTape for automatic differentiation.
with tf.GradientTape() as g:
reconstructed_image = decoder(encoder(x))
loss = mean_square(reconstructed_image, x)

# Variables to update, i.e. trainable variables.
trainable_variables = weights.values() + biases.values()

# Compute gradients.
gradients = g.gradient(loss, trainable_variables)

# Update W and b following gradients.
optimizer.apply_gradients(zip(gradients, trainable_variables))

return loss`

the line:

trainable_variables = weights.values() + biases.values()

results to me in an error due to the impossibility of summing two python dict_values. I personally solved the issue by converting both dict_values to lists:

trainable_variables = list(weights.values()) + list(biases.values())

I hope my issue is of utility for this great repo! Thank you for the work

[xy-always]

已经收到您的邮件。

[xmingpyccnu]

您好！很开心收到您的邮件。我将在查看邮件后，尽快给您回复。

[gpendrys]

Your comment helped me. Thank you!

[xy-always]

已经收到您的邮件。

[Alexzhang369]

这是来自QQ邮箱的假期自动回复邮件。您好，您的邮件已经收到。谢谢您，祝顺利。

[jiandandema]

谢谢您给马坤发送邮件，我会及时查看的，谢谢，辛苦了