@tranquil dirge has leveled up! (0 ➜ 1)
#Help understanding tensor shape inputs and outputs
desert wagonBOT
restive sluice
it looks like you are providing a set of 3 by 1 values. a single sample input in your case should be only [3]. but for the model, you need to provide a batch of [something, 3]
restive sluice
sure, np. glad you got it figured out
restive sluice
sigmoid can only return values between 0 and 1, if your values are exactly 0/1, your model is either very sure, or maybe the optimizer is exploding your gradients (maybe try a default adam optimizer?)
you should only need to train once on a dataset that includes enough data to generalize for your cases, then you can predict as much as you want