如何为预训练的 LSTM 模型提供一个样本文本输入

Question

I am trying to do toxic comment classification. I found a dataset in https://www.kaggle.com/c/jigsaw-toxic-comment-classification-challenge . It has many comments with corresponding values for label class "toxic", "severe_toxic" ..etc. I want to classify my single text input to the corresponding labeled class. I have created and trained a model using LSTM. Now I want to give a single text sentence as input to the model to predict the output. But I don't know how to convert the text input and give it to the trained model.

The source-code has been obtained from kaggle challange .

summary of the model

Answer 1

Prediction can be done using predict() function as below:

y_predict = model.predict(X_te, batch_size=batch_size) Where, X_te is the pre-processed test-set. The pre-processing is generally same for training-set and test-set.

In case, if you want to predict for a single instance from the test set, the input has to be reshaped, as given below:

y_pred = model.predict(X_te[0].reshape(200,))