Error in TFLearn model for CNN text analysis

Question

I've tried to use Karan's model TFlearn model "as is" as a template for CNN model for sentiment analysis, but I've got 2 strange errors:

1) there is no module global_max_pool in tflearn.layers.conv package (i've tried two tflearn packages with the same result, so I just used source code of it (all other modules were imported w/o problems).

2) ----> 5 branch1 = conv_1d(network, 128, 3, padding='valid', activation='relu', regularizer="L2") AssertionError: Incoming Tensor shape must be 3-D

Do you have any idea what is the reason? how I can overcome it?

Answer 1

To help others, the error occurs due to old versions of TensorFlow and TFLearn. Please update to the latest version using:

pip install --upgrade tensorflow
pip install --upgrade tflearn