Artificial Intelligence Stack Exchange
Artificial Intelligence Stack Exchange

Questions tagged [validation-loss]

For questions about the validation loss, i.e. the loss of a model (e.g. a neural network) computed on the validation data (or dataset), which is different from the testing/test data (i.e. the data that you use after having trained your model(s) and selected the best model according to the validation loss/performance).

9 questions
2
votes
1 answer

Fluctuations in loss during in epoch evaluation of GRU

I am training a one-layer unidirectional vanilla GRU on a next item prediction task with regard to the last 10 interacted items. In my original experiment, where I trained on approx. 5.5M samples and validated on around 1M samples, I saw periodic…
PatrickSVM
  • 53
  • 3
1
vote
0 answers

React on train-validation curve after trening

I have a regression task that I tray to solve with AI. I have around 6M rows with about 30 columns. (originally there was 100, but I reduce it with drop feature importance) I understand basic principle: Look if model overfit or underfit - according…
Marko Zadravec
  • 169
  • 8
0
votes
0 answers

MSE Loss for human pose keypoints detection heatmap regression

The Gaussian function used to create heatmap with Gaussian peak at the ground truth coordinates. And in the training loop the distance between ground truth training data and model predicted heatmap calculated using MSE loss function. MSE = (1/n) *…
Ben Goz
  • 1
0
votes
0 answers

Is It Normal for Validation Accuracy to be Higher than Training Accuracy?

I'm training a deep learning model in PyTorch for a classification problem, and I’ve noticed that the validation accuracy is consistently higher than the training accuracy throughout the training process. This behavior persists even after tuning the…
Kat27
  • 1
0
votes
1 answer

Validation loss is always lower than training loss whatever i try

I've been training several types of MLPs with different optimisers and tuned them with keras's hyperband tuner. All of them follow this cone architecture: All the networks were trained on the same dataset composed of 342k input examples and another…
Roger Smith
  • 1
  • 1
0
votes
1 answer

What are possible reasons for the validation loss increasing with more data?

I trained a neural network on an NLP problem and compared the loss and BLEU score on the validation data with the same training parameters in two scenarios: a) when I trained on 25% of the data, b) when I trained on 100% of the data. I observed a…
postnubilaphoebus
  • 356
  • 2
  • 13
0
votes
1 answer

What would be a good cost function based on both saliency-maps and labels?

I have a number of input samples where: every input sample has both a label and a reference-map. This reference-map gives a score to each location of an input sample. The score defines how much this location of the input sample SHOULD contribute to…
Wtt
  • 1
  • 1
0
votes
0 answers

What can cause massive instability in validation loss?

I'm working with very weird data that is apparently very hard to fit. And I've noticed a very strange phenomenon where it can go from roughly 0.0176 validation MSE to 1534863.6250 validation MSE in only 1 epoch! It usually then will return to a very…
profPlum
  • 496
  • 2
  • 10
0
votes
1 answer

Is it okay to calculate the validation loss over batches instead of the whole validation set for speed purposes?

I have about 2000 items in my validation set, would it be reasonable to calculate the loss/error after each epoch on just a subset instead of the whole set, if calculating the whole dataset is very slow? Would taking random mini-batches to calculate…
Ilknur Mustafa
  • 115
  • 4