1D Regression

Controls

Learning Rate: 0.01

Training Speed

Optimizer

Network Architecture

Input Layer (1 neuron)

x-coordinate (single number)

Hidden Layer 1 (10 neurons)

Dense layer with ReLU activation

Transforms input into 10 features

Hidden Layer 2 (10 neurons)

Dense layer with ReLU activation

Further processes the 10 features

Output Layer (1 neuron)

Outputs: predicted y-coordinate

Training Status:

Training Points: 0

Epochs Completed: 0

Current Loss: 0.000000

Total MSE (All Points): 0.000000

Optimizer: ADAM

Status: Stopped

Speed: Slow

Training

Durign training, the network is;
For each epoch:
  For each training point (x, y):
    1. Feed x to input neuron
    2. Get prediction ŷ from output neuron
    3. Calculate MSE = (ŷ - y)²
    4. Adjust network weights to reduce this error
    5. Move to next point

Click on graph to add training points

W & B

Weights & Biases will be here after training has begun