models/efficientnet_v2_l.py [115:185]: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - self.features_21_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_22_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_22_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_22_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_22_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_22_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_22_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_22_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_22_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_22_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_22_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_23_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_23_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_23_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_23_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_23_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_23_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_23_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_23_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_23_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_23_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_24_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_24_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_24_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_24_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_24_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_24_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_24_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_24_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_24_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_24_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_25_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_25_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_25_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_25_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_25_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_25_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_25_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_25_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_25_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_25_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_26_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_26_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_26_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_26_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_26_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_26_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_26_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_26_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_26_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_26_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_27_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_27_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_27_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_27_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_27_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_27_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_27_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_27_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_27_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_27_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_28_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_28_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_28_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_28_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_28_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_28_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_28_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_28_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_28_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_28_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - models/efficientnet_v2_xl.py [103:173]: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - self.features_21_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_22_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_22_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_22_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_22_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_22_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_22_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_22_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_22_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_22_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_22_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_23_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_23_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_23_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_23_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_23_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_23_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_23_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_23_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_23_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_23_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_24_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_24_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_24_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_24_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_24_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_24_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_24_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_24_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_24_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_24_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_25_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_25_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_25_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_25_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_25_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_25_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_25_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_25_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_25_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_25_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_26_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_26_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_26_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_26_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_26_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_26_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_26_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_26_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_26_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_26_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_27_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_27_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_27_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_27_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_27_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_27_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_27_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_27_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_27_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_27_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) self.features_28_conv_0 = torch.nn.modules.conv.Conv2d(192, 768, 1, 1, 0, bias=False) self.features_28_conv_1 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_28_conv_3 = torch.nn.modules.conv.Conv2d(768, 768, 3, 1, 1, groups=768, bias=False) self.features_28_conv_4 = torch.nn.modules.batchnorm.BatchNorm2d(768) self.features_28_conv_6_avg_pool = torch.nn.modules.pooling.AdaptiveAvgPool2d(1) self.features_28_conv_6_fc_0 = torch.nn.modules.linear.Linear(768, 48) self.features_28_conv_6_fc_2 = torch.nn.modules.linear.Linear(48, 768) self.features_28_conv_6_fc_3 = torch.nn.modules.activation.Sigmoid() self.features_28_conv_7 = torch.nn.modules.conv.Conv2d(768, 192, 1, 1, 0, bias=False) self.features_28_conv_8 = torch.nn.modules.batchnorm.BatchNorm2d(192) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -