All translations

Enter a message name below to show all available translations.

Found 3 translations.

Name	Current message text
^h English (en)	Training an RNN requires computing gradients of the loss with respect to the shared weights. '''{{Term\|backpropagation}} through time''' (BPTT) "unrolls" the RNN across time {{Term\|training step\|steps}}, producing a deep {{Term\|feedforward neural network\|feedforward network}} with shared weights, and then applies standard [[Backpropagation\|backpropagation]].
^h Spanish (es)	Entrenar una RNN requiere calcular gradientes de la pérdida con respecto a los pesos compartidos. La '''{{Term\|backpropagation\|retropropagación}} a través del tiempo''' (BPTT) "despliega" la RNN a lo largo de los {{Term\|training step\|pasos}} de tiempo, produciendo una {{Term\|feedforward neural network\|red feedforward}} profunda con pesos compartidos, y luego aplica la [[Backpropagation\|retropropagación]] estándar.
^h Chinese (zh)	训练 RNN 需要计算损失相对于共享权重的梯度。'''跨时间{{Term\|backpropagation\|反向传播}}'''（BPTT）将 RNN 沿时间{{Term\|training step\|步}}"展开"，生成一个权重共享的深层{{Term\|feedforward neural network\|前馈网络}}，然后应用标准的[[Backpropagation\|反向传播]]。