Normally any deep learning network having only forward method and same number of inputs during training and inference also but your uni planner network having 7 inputs during training, but only 4 input during inference
And why uni planner network having infer method also and how it trained with 4 inputs without loss function
could you please explain it?
Normally any deep learning network having only forward method and same number of inputs during training and inference also but your uni planner network having 7 inputs during training, but only 4 input during inference And why uni planner network having infer method also and how it trained with 4 inputs without loss function could you please explain it?