Test Validated Field with airSlate SignNow

Test validated field

[Music] in this video we'll be discussing the different data sets used during training and testing and neural network for training and testing purposes for our model we should have our data broken down into three distinct data sets these data sets will consist of a training set a validation set and a test set let's start by discussing the training set the training set is what it sounds like it's the set of data used to train the model during each epoch our model will be trained over and over again on this same data in our training set and it will continue to learn about the features of this data the hope with this is that later we can deploy our model and have it accurately predict on new data that it's never seen before it'll be making these predictions based on what it learned about the training data okay now let's discuss the validation set the validation set is a set of data separate from the training set that is used to validate our model during training this validation process helps give information that may assist us with adjusting our hyper parameters so recall how we just mentioned that with each epoch during training the model will be trained on the data in the training set well it will also be simultaneously validating on the data in the validation set now we know from our previous videos on training that during the training process the model will be classifying the output for each input in the training set after this classification occurs the loss will then be calculated and the weights in the model will be adjusted then during the next epoch it will classify the same input again now also during training the model will be classifying each input from the validation set as well it will be doing this classification based only on what it's learned about the data it's being trained on in the training set the weights won't be updated in the model based on the loss calculated from our validation data remember the data and the validation set is separate from the data in the training set so when it's validating on this data this data does not consist of samples that the model is already familiar with from training one of the major reasons we need a validation set is to ensure that our model is not overfitting to the data in the training set we'll discuss overfitting and underfitting in detail at a later time but the idea of overfitting is that our model becomes really good at being able to classify the data in the training set it's unable to generalize and make accurate classifications on data that it wasn't trained on so during training if we're also validating the model on our validation set and see that the results it's giving for the validation set are just as good as the results it's giving for the training data then we can be more confident that our model is not overfitting on the other hand if the results on the training data are really good but the results on the validation data are lagging behind then our model is likely overfitting now let's move on to the test set the test set is a set of data that's used to test the model after the model has already been trained this test set is separate from both the training set and validation set after our model has been trained and validated using our training and validation sets we'll then use our model to predict the output of the data in the test set now one major difference between the test set and the other two sets that we've discussed is that the test set should not be labeled the training set and validation set has to be labeled so that we can see the metrics given during trainings like the loss in the accuracy from each epoch so when the model is predicting on unlabeled data in our test set this would be the same type of process that would be used if we were to deploy our model into the field for example if in practice we're using a model to classify data without knowing what the labels of the data are beforehand or would never have being shown the exact date it's going to be classifying then of course we wouldn't be giving our model the labeled data to do this the entire goal of having a model be able to classify is to do this without knowing what the data is beforehand now hopefully we have an idea about how our data should be organized let's take a look now at how to structure our data for training our model with Kerris in a previous video we trained at this model shown here using the Charis fit function now I'm just going to copy this fit function down to a clean space in our notebook now whenever we fit the model I said in a previous video that we are passing the train Sables and train labels as the first two parameters to the fit function now these train samples are our entire training set and caris is expecting this training set to be in a format of either a numpy array or a list of dump I erase so I'm just going to print out this training set here so that you can see that it's just simply numpy right here now the training labels Kerris is expecting those to be in the format of a number array so again we'll just print this out here as well I'm not going to focus on this underlying data since we're just illustrating the concept here as mentioned when we use this data before I will pop up the video on the screen so that you can see in another video what this data actually is and how we pre processed it for this training our training set and our training labels there are the first two parameters passed to the fit function whenever we're training our model now what about our validation set well you don't even have to explicitly make a validation set you can pass in the parameter validation split and give it a fraction that instructs Kharis to split out this fraction of data from your training set and use it as your validation set I just arbitrarily chose 20% here so it's going to go into my training samples extract 20% out from training and use it only for validation so now if we run this fit function recall before that we only had the loss and accuracy metrics now since we added a validation set we also have the validation loss and validation accuracy metrics as well now one other thing to point out is for our validation set we're kind of implicitly creating this validation set here using the validation split parameter but we could also create an explicit validation set that is separate from the training set altogether and then passed that entire set into our model as well and to do that we wouldn't be using this validation split parameter instead if we scroll down we see we would be using this validation data parameter instead and passing in the validation set explicitly here now I don't actually have the validation set created but this is the format that Kerris would expect it would expect a list of tuples and each tuple would be the actual sample the actual data point and the label so here I've just illustrated that within this list we have a sample and a label and then another sample and another label and that would make up the entire list okay so those are our two options that we have to work with for our validation set with Kerris now the test set would be structured just like the training set is and we'd be using the test set when we call the predict function on our model since we haven't yet covered the concept of predicting all saves showing how to do that with Kerris until we cover the concept in another video so hopefully now you have an idea about what the different data sets are for training and testing and neural network and how you can make use of them in Carris and I hope you found this video helpful if you did please like the video subscribe suggest and comment and thanks for watching [Music]