search

Chancing0 / EZ-LSTM-Speech-Recognition-Project

An ez LSTM Speech Reconization project for beginner
7 stars 2 forks source link

What is the difference between Validation data and Training Data #1

Open AiJunXian opened 6 years ago

AiJunXian commented 6 years ago

I have the dataset from the original project.. I see that you created two paths:

tra_data_path=r"D:\python_code\tensorflow_works\train_10\pre" val_data_path=r"D:\python_code\tensorflow_works\valx\pre"

I put same data in both paths? or I choose different data?

Chancing0 commented 6 years ago

I'm so sorry , I haven't use my own accunt half years, i'm a student, and a beginner.I also very happy if i can help you . tra_data_path is the Path of your trainningset val_data_path is the Path of your validationset In general , your total dataset should be separate as 60%(traning ), 20%(validationset),20% (testing). (produces a 60%, 20%, 20% split for training, validation and test sets.) so you should choose different data . for more information ,check in https://www.quora.com/Should-I-split-my-data-to-train-test-split-or-train-validation-test-subset it maybe too late to reply your question , sorry again.