Hi - I could not understand how the UNITS=3 in this for Layer 1. Given there is only one example of 2 features (200 Celcius, 17min)

How does one set UNITS count? What is the logic

Hi - I could not understand how the UNITS=3 in this for Layer 1. Given there is only one example of 2 features (200 Celcius, 17min)

How does one set UNITS count? What is the logic

What is the time mark for your question in the lecture?

Any update ?

Not yet, I do not have access to the lectures at the moment since Coursera updated their Terms of Use.

@Zerxes24, the “units” parameter refers to the number of units (nodes) in the layer. So, in this case, we’ve chosen to have 3 nodes in layer 1. This is represented by the 3 circles in the diagram for layer 1

1 Like

What is happening here is that first we are initialisation a single example that is x(200,17) which is then taken as input by 3 separate logistic regression units. The 3 units in the layer are 3 separate logistic regression models and they have their own weights and biases. I guess u got confused thinking that the number of inputs should be equal to the number of units in the next later.

1 Like

Hi, the `units=3`

parameter in the first layer means there are 3 neurons in that layer. This number is not related to the number of input features (which are temperature and duration) but is a hyperparameter you choose to help the model learn better. More neurons can capture more complex patterns, but it’s a balance between complexity and performance. We often start with a small number of neurons and adjust based on how well the model performs. In this case, using 3 neurons is a starting point to allow the network to process and learn from the input data effectively.

To be clear, what you say is true only for the hidden layers.

1 Like