# Network Input Format


📗 An example fully connected network:


➩ Layers are separated by five dashes "-----" or more.
➩ Weights from each unit in previous layer are separated by new line "\n". Biases in the last line.
➩ Weights connected to each unit in the next layer are separated by comma ",".



📗 Number of units:

📗 Activation:


📗 Input:
📗 Output:
📗 Tensorflow Output (should be the same):

# Train Small Networks


📗 Input training data: y is the first column

➩ Or separate y (if not first or last column):


📗 Number of epochs:

📗 Learning rate (ADAM):

📗 Batch size (1 for SGD):


➩ Status:


📗 Trained network will be put at the top of the page.

# Generate Random Training Sets


📗 Number of features:
from to
📗 Number of samples:

📗 Function to generate y (written in JavaScript): y = function(x) {...}, for example, return x[0] > x[1] ? 0 : (x[1] > x[2] ? 1 : 2);.







Last Updated: April 23, 2026 at 2:39 AM