Use inverse transform with deep learning Conceptual clarity needed

Question

I have built a deep learning model that encoded the value using LabelEncoder() which is further being converted to categorical values using np_utils.to_categorical()

labels = np.array(labels)
transformed_features = docs_encoded.astype(int)
encoder = LabelEncoder()
encoder.fit(labels)
encoded_labels = encoder.transform(labels)
transformed_labels = np_utils.to_categorical(encoded_labels)

Predicting gives me results like : array([[2.40659632e-04, 3.97350988e-04, 4.11552377e-04, ...,

Doing np.argmax(to_categorical(encoder.inverse_transform(encoded_value))) rasies a ValueError: invalid literal for int() with base 10: 'coldstorage'. I know it just means that the argument might be something other than a digit/number. Can't figure this one out. Need help with converting the numbers back to user_ids. Thanks.

My data looks like:

user_id     date                body

mahamudra   2015-01-14T16:52:11 Hi, just search for mahamudra in Evolution.    
mahamudra   2015-01-15T13:55:38 Thank you!  Do I have to register again as vendor?  I tried but it said: 
Shart       2015-02-25T18:43:25 what gives?  is it true missy took the coin and run?
Shart       2015-02-25T18:42:24 people were warned not to use panacea. but no worries,misssy will find a good way to spend your coins.

Comments

Hey @Anan,

Can you please provide the code you have written for this? Or if you have used 'int' anywhere in your code, just use 'float' instead of that. I hope your error will go off.

To learn more, visit our deep learning online course