Closed padmaja9 closed 6 years ago
Hi @padmaja9 , Thanks for the message!
I haven't gotten around to adding an evaluation part for new data at the end of the notebook. It's a great idea and I will see if I can do that this weekend.
If you can't wait, here's what it would entail:
eval_data = clean_text(eval_text)
function.vocab_processor
like the line: eval_processed = np.array(list(vocab_processor.fit_transform(eval_data)))
that is in the code.output_logits
, which are the model output values before the softmax via something like: eval_output = sess.run(logits_out, feed_dict={x_data: eval_processed, dropout_keep_prob: 1.0})
Then the prediction is the np.argmax(eval_output)
. I'll see if I can update the jupyter notebook this weekend.
Hi @nfmcclure , Thanks for the response. I added the evaluation part to my code but i'm unable to get the prediction correctly.It would be great if you could add this evaluation part this weekend.
Hi @padmaja9 , Please checkout the bottom of the jupyter notebook: https://github.com/nfmcclure/tensorflow_cookbook/blob/master/09_Recurrent_Neural_Networks/02_Implementing_RNN_for_Spam_Prediction/02_implementing_rnn.ipynb
I've added some code to do predictions on new texts using the model that is in memory. I hope that helps.
I'm going to close this issue. If you have further questions, feel free to re open.
I'm really a beginner with tensor flow and in all of this field.i was asked to do a spam detection using Recurrent Neural Network.I took a reference of your code. I went through this code https://github.com/nfmcclure/tensorflow_cookbook/blob/master/09_Recurrent_Neural_Networks/02_Implementing_RNN_for_Spam_Prediction/02_implementing_rnn.ipynb it is working fine.but i didn't understand where to give our own message as input to predict whether it is spam or ham. could you please help me out in finding the solution for this