Hmm. Okay. Could I ask why these models are very sensitive to prior initialization? I’m not the most experienced nor skilled in this type of models.
Do these types of networks need input features to be a certain format or scale?
Hmm. Okay. Could I ask why these models are very sensitive to prior initialization? I’m not the most experienced nor skilled in this type of models.
Do these types of networks need input features to be a certain format or scale?