Figure 3 | Scientific Reports

Figure 3

From: A scalable convolutional neural network approach to fluid flow prediction in complex environments

Figure 3

Five-fold cross-validation of the single domain model. Loss (error) in relation to the validation data is shown in red, whereas loss associated with the test datasets is shown in blue. Error bars represent standard deviations of loss for each finite set of CFD data chosen. (Inset) Five-fold cross-validation of the single domain model where the y-axis (loss) is in log scale. Straight (dotted) line in the logarithm scale indicates exponential decay, provided for reference.

Back to article page