Can anyone please suggest me how to predict the tags make into readable format after obtaining the results from mode.predict. The words have been converted into word embedding and numpy arrays and then trained and tested. how get this back to original readable format?
You have used the dictionary list to convert the words to a one-hot encoded sequence right,? Just use the argmax value to find the index in the list and hence the word.
if any doubts contact me at:
It depends on the number of nodes in the output layer if you use the number of classes then you can use argmax function else if you use only one neuron then you set a threshold to obtain your predicted class.
https://www.tensorflow.org/api_docs/python/tf/math/argmax
Labels can be saved in python Dict and when you get the label from the model just search for it in Dict and you will get Label again.
you can also use list, but I will suggest using Dict.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python