All translations
Enter a message name below to show all available translations.
Found 3 translations.
| Name | Current message text |
|---|---|
| h English (en) | {| class="wikitable" |- ! Method !! Normalizes over !! Use case |- | '''Batch Norm''' || Batch and spatial dims, per {{Term|feature map|channel}} || {{Term|convolutional neural network|CNNs}} with large batches |- | '''Layer Norm''' || All {{Term|feature map|channels}} and spatial dims, per sample || {{Term|transformer|Transformers}}, {{Term|recurrent neural network|RNNs}}, small batches |- | '''Instance Norm''' || Spatial dims only, per sample per {{Term|feature map|channel}} || Style transfer, image generation |- | '''Group Norm''' || Groups of {{Term|feature map|channels}}, per sample || {{Term|object detection}}, small-batch training |} |
| h Spanish (es) | {| class="wikitable" |- ! Método !! Normaliza sobre !! Caso de uso |- | '''Batch Norm''' || Dimensiones de lote y espaciales, por {{Term|feature map|canal}} || {{Term|convolutional neural network|CNN}} con lotes grandes |- | '''Layer Norm''' || Todos los {{Term|feature map|canales}} y dimensiones espaciales, por muestra || {{Term|transformer|Transformers}}, {{Term|recurrent neural network|RNN}}, lotes pequeños |- | '''Instance Norm''' || Solo dimensiones espaciales, por muestra y por {{Term|feature map|canal}} || Transferencia de estilo, generación de imágenes |- | '''Group Norm''' || Grupos de {{Term|feature map|canales}}, por muestra || {{Term|object detection|Detección de objetos}}, entrenamiento con lotes pequeños |} |
| h Chinese (zh) | {| class="wikitable" |- ! 方法 !! 归一化范围 !! 适用场景 |- | '''Batch Norm''' || 批量维度和空间维度，按{{Term|feature map|通道}} || 大批量的{{Term|convolutional neural network|CNN}} |- | '''Layer Norm''' || 所有{{Term|feature map|通道}}和空间维度，按样本 || {{Term|transformer|Transformer}}、{{Term|recurrent neural network|RNN}}、小批量场景 |- | '''Instance Norm''' || 仅空间维度，按样本按{{Term|feature map|通道}} || 风格迁移、图像生成 |- | '''Group Norm''' || {{Term|feature map|通道}}分组，按样本 || {{Term|object detection|目标检测}}、小批量训练 |} |