towhee.models.layers¶

`towhee.models.layers.activations`
`towhee.models.layers.aspp`
`towhee.models.layers.attention`
`towhee.models.layers.cond_conv2d`
`towhee.models.layers.conv2d_same`
`towhee.models.layers.conv2d_separable`
`towhee.models.layers.conv4d`
`towhee.models.layers.conv_bn_activation`
`towhee.models.layers.convmlp`
`towhee.models.layers.cross_attention`
`towhee.models.layers.dropblock2d`	PyTorch implementations of DropBlock and DropPath (Stochastic Depth) regularization layers.
`towhee.models.layers.droppath`	DropPath
`towhee.models.layers.ffn`
`towhee.models.layers.gatedmlp`
`towhee.models.layers.layers_with_relprop`
`towhee.models.layers.mixed_conv2d`
`towhee.models.layers.mlp`
`towhee.models.layers.multi_scale_attention`
`towhee.models.layers.multi_scale_transformer_block`
`towhee.models.layers.netvlad`
`towhee.models.layers.non_local`
`towhee.models.layers.padding_functions`
`towhee.models.layers.patch_embed2d`
`towhee.models.layers.patch_embed3d`
`towhee.models.layers.patch_merging`
`towhee.models.layers.patch_merging3d`
`towhee.models.layers.pool_attention`
`towhee.models.layers.position_encoding`	Various positional encodings for the transformer.
`towhee.models.layers.relative_self_attention`
`towhee.models.layers.resnet_basic_3d_module`
`towhee.models.layers.sam`
`towhee.models.layers.sequence_pool`
`towhee.models.layers.spatial_temporal_cls_positional_encoding`
`towhee.models.layers.spp`
`towhee.models.layers.swin_transformer_block3d`
`towhee.models.layers.temporal_cg_avgpool3d`
`towhee.models.layers.tf_avgpool3d`
`towhee.models.layers.time2vec`
`towhee.models.layers.transformer_encoder`
`towhee.models.layers.vision_transformer_basic_head`
`towhee.models.layers.window_attention`
`towhee.models.layers.window_attention3d`