Questions tagged [training]

The tag has no usage guidance.

Filter by
Sorted by
Tagged with
33
votes
8answers
7k views

What would I prefer - an over-fitted model or a less accurate model?

Let's say we have two models trained. And let's say we are looking for good accuracy. The first has an accuracy of 100% on training set and 84% on test set. Clearly over-fitted. The second has an ...
94
votes
2answers
70k views

Training an RNN with examples of different lengths in Keras

I am trying to get started learning about RNNs and I'm using Keras. I understand the basic premise of vanilla RNN and LSTM layers, but I'm having trouble understanding a certain technical point for ...
31
votes
4answers
13k views

Is it always better to use the whole dataset to train the final model?

A common technique after training, validating and testing the Machine Learning model of preference is to use the complete dataset, including the testing subset, to train a final model to deploy it on, ...
26
votes
2answers
11k views

What is the advantage of keeping batch size a power of 2?

While training models in machine learning, why is it sometimes advantageous to keep the batch size to a power of 2? I thought it would be best to use a size that is the largest fit in your GPU memory /...
14
votes
4answers
41k views

Train, test split of unbalanced dataset classification

I have a model that does binary classification. My dataset is highly unbalanced, so I thought that I should balance it by undersampling before I train the model. So balance the dataset and then ...
7
votes
2answers
2k views

How to fix class imbalance in training sample?

I was very recently asked in a job interview about solutions to fix an imbalance of classes in the training dataset. Let's focus on a binary classification case. I offered two solutions: oversampling ...
44
votes
5answers
35k views

Should a model be re-trained if new observations are available?

So, I have not been able to find any literature on this subject but it seems like something worth giving a thought: What are the best practices in model training and optimization if new observations ...
7
votes
3answers
4k views

Build a tool for manually classifying training data images

I have a large number of images that I need to classify for training a clustering algorithm, and I would like to do so offline (the data is proprietary). Basically, I'd like to build a desktop survey ...
9
votes
1answer
7k views

How to train data by batch from disk?

I am working on a convolutional neural network for image classification. The training dataset is too large to be loaded on my computer memory (4gb), on top of that I also need to try some augmentation ...
6
votes
4answers
12k views

Tool to Label Images for Supervised Classification

I have a couple thousand photos of whales taken from drones and I'm planning to build a simple binary classifier to run on these and future images to see if they contain a whale. I'd like to label ...
4
votes
1answer
944 views

“other” class in Image classification

In the MNIST dataset, you have 10 defined classes, one for each digit. But you don't have a "not a digit" class. It seems that most image classification datasets are the same. But in a business ...
4
votes
1answer
231 views

How to predict based on multiple samples?

I am relatively new to ML so I apologies in advance if my question shows lack of understating of the field. The problem A particular study course has a high drop-out rate and we want to reduce it. ...
3
votes
1answer
142 views

When forecasting time series, how does one incorporate the test data back into the model after training?

When you build a classification or regression model, you typically split the data into a train data set and a test data set. The test data is a randomly selected subset of the overall data. Once you ...
2
votes
0answers
151 views

How to resolve the instability of average reward per episode in training of DQN (Deep Q-Network)?

what is shown when average reward per episode in training is unstable? If there is big difference between average reward per episode and final reward by test section, what we can say? For instance in ...
8
votes
1answer
1k views

CNN for phoneme recognition

I am currently studying this paper, in which CNN is applied for phoneme recognition using visual representation of log mel filter banks, and limited weight sharing scheme. The visualisation of log ...
3
votes
2answers
654 views

Why can't I choose my hyper-parameter in the training set?

Say I've divided the data into 3 parts: training, validation and test. I know for example, that in Neural Networks, the number of hidden layers is a hyper parameter. Why can't I train numerous NN ...
3
votes
2answers
65 views

Can a classifier be trained with reinforcement learning without access to single classification results?

Question: Can a classifier be trained with reinforcement learning without access to single classification results? I want to train a classifier using reinforcement learning. However, there is one big ...
2
votes
1answer
27 views

what does the standard deviation plot around my learning curve indicate?

I plotted a learning curve below. There is a thick red band around the top portion of my training score. Why is it so high at the beginning? Below is a snippet of the code used: ...
2
votes
2answers
897 views

DC GAN with Batch Normalization not working

I'm trying to implement DC GAN as they have described in the paper. Specifically, they mention the below points Use strided convolutions instead of pooling or upsampling layers. Use only one fully ...
1
vote
2answers
1k views

validation/training accuracy and overfitting

If we randomly split the data into training data and validation data, and assume the training data and validation data have similar "distributions", i.e. they are both good representations of the ...
0
votes
2answers
28 views

Present results of the best or the last iteration on dev set?

Which is the correct way - presenting the results of the best or the last iteration on the dev set in a paper? In research papers I usually see only one value, is it the best iteration of all? I'm ...