Квизы: Neural Networks & Deep Learning

How neural networks really work: neurons and activations, losses, backpropagation, CNNs and the tricks that make deep networks train.

Средний Квизов: 5 · вопросов: 50 около 33 минуты
0/5сдано Экзамен по разделу 20 вопросов вперемешку. За 80% - сертификат.
Теги:
1 Neurons & Activation Functions Лёгкий Learn what a single neuron computes and how ReLU, sigmoid, tanh and softmax differ. #neural-networks#activations 10 вопр. · ~5 мин Не начат Начать
2 Forward Pass & Loss Functions Средний Follow data through a layer, count parameters and compute softmax and cross-entropy by hand. #neural-networks#loss 10 вопр. · ~6 мин Не начат Начать
3 Backpropagation & Gradient Descent Сложный Apply the chain rule, run gradient descent steps by hand and explain why training succeeds or fails. #backpropagation#gradient-descent#interview 10 вопр. · ~8 мин Не начат Начать
4 Convolutional Neural Networks (CNNs) Средний Calculate convolution output sizes and parameter counts and understand pooling and padding. #cnn#computer-vision 10 вопр. · ~6 мин Не начат Начать
5 Training Deep Networks Сложный Diagnose and fix real training problems with learning rates, dropout, batch norm and early stopping. #regularization#optimization 10 вопр. · ~8 мин Не начат Начать