Pergunta de entrevista da empresa LandingAI
Why do backpropagation instead of calculating gradients numerically?
Resposta da entrevista
Sigiloso
23 de abr. de 2019
Caching makes backprop highly efficient due to the chain rule.
Why do backpropagation instead of calculating gradients numerically?
Sigiloso
Caching makes backprop highly efficient due to the chain rule.