I have resized my data images using the “Image Resizer” node to have same size of 500,500,1 .
Now i get the fellowing error from the node “DL4J Feedforward Learner (Classification)”
ERROR DL4J Feedforward Learner (Classification) 2:36:31 Execute failed: Invalid input: expect output columns must be equal to rows 500 x columns 500 x channels 3 but was instead [128, 250000]
Please your help
What’s the size of your input data? It expects 500x500x3 sized images.
According to the node "Image Properties " it is 500,500,1
As the error message states, the network expects 3 input channels.
Now i have changed the Data Parameter -> Image size to be 500,500,1.
The Knime shut down when it reaches the max heap memory. I have 15 GB RAM memory.
Can i deal with memory problems?
Did you increase heapspace? Also in the DL4J preference you can change the amount of memory DL4J can actually use (e.g. to 8000m).