2. Liga Interregional Group 2 stats & predictions

What makes Group 2 unique compared to other leagues?

Group 2 stands out due to its competitive nature, diverse playing styles, and strong community involvement. It serves as a breeding ground for future football stars while offering thrilling entertainment for fans. [0]: import os [1]: import random [2]: import pickle [3]: import numpy as np [4]: import torch [5]: import torch.nn as nn [6]: import torch.nn.functional as F [7]: from collections import OrderedDict [8]: from PIL import Image [9]: def count_parameters(model): [10]: return sum(p.numel() for p in model.parameters() if p.requires_grad) [11]: def weights_init(m): [12]: classname = m.__class__.__name__ [13]: if classname.find('Conv') != -1: [14]: m.weight.data.normal_(0.0, 0.02) [15]: elif classname.find('BatchNorm') != -1: [16]: m.weight.data.normal_(1.0, 0.02) [17]: m.bias.data.fill_(0) [18]: class GaussianNoise(nn.Module): [19]: def __init__(self): [20]: super(GaussianNoise,self).__init__() [21]: def forward(self,x): [22]: noise = torch.randn_like(x)*0.1 [23]: return x+noise [24]: class ResidualBlock(nn.Module): [25]: def __init__(self,in_features,out_features,batch_norm=True): [26]: super(ResidualBlock,self).__init__() [27]: conv_block = [] [28]: conv_block += [nn.ConvTranspose1d(in_features,out_features, [29]: kernel_size=3,stride=1,padding=1,bias=False)] [30]: if batch_norm: [31]: conv_block += [nn.BatchNorm1d(out_features)] self.conv_block = nn.Sequential(*conv_block) self.relu = nn.ReLU() self.skip_proj = nn.ConvTranspose1d(in_features,out_features, kernel_size=1,stride=1,padding=0,bias=False) self.out_channels = out_features self.model = nn.Sequential( OrderedDict([('conv_block', self.conv_block), ('relu',self.relu)])) self.residual_function = nn.Sequential( OrderedDict([('conv_block', self.conv_block), ('relu',self.relu), ('skip_proj',self.skip_proj)])) self.identity_function = nn.Identity() def forward(self,x): return self.model(x) + self.residual_function(x) __all__ = ['ResidualBlock']<|repo_name|>zhaohaoxu/DAE<|file_sep|>/model/resnet.py import os import random import pickle import numpy as np import torch import torch.nn as nn import torch.nn.functional as F from collections import OrderedDict from PIL import Image def count_parameters(model): return sum(p.numel() for p in model.parameters() if p.requires_grad) def weights_init(m): classname = m.__class__.__name__ if classname.find('Conv') != -1: m.weight.data.normal_(0.0, 0.02) elif classname.find('BatchNorm') != -1: m.weight.data.normal_(1.0, 0.02) m.bias.data.fill_(0) class GaussianNoise(nn.Module): def __init__(self): super(GaussianNoise,self).__init__() def forward(self,x): noise = torch.randn_like(x)*0.1 return x+noise class ResNet(nn.Module): def __init__(self,in_channels,out_channels, n_res_blocks,n_downsampling_blocks,n_upsampling_blocks, residual_channels,growth_channels, lr_mult=1): super(ResNet,self).__init__() self.n_downsampling_blocks = n_downsampling_blocks self.n_upsampling_blocks = n_upsampling_blocks self.in_channels = in_channels self.out_channels = out_channels self.residual_channels = residual_channels self.growth_channels = growth_channels self.lr_mult = lr_mult self.downsampling_layers = nn.ModuleList() for i in range(n_downsampling_blocks): residual_in_channels = residual_in_channels*2 if i!=0 else residual_in_channels residual_out_channels = residual_out_channels*2 if i!=n_downsampling_blocks-1 else out_channels downsample_layer = DownsampleLayer(residual_in_channels,residual_out_channels,growth_channels) self.downsampling_layers.append(downsample_layer) self.residual_layers = nn.ModuleList() for i in range(n_res_blocks+n_downsampling_blocks+n_upsampling_blocks): residual_in_channels = residual_in_channels*2 if i=n_downsampling_blocks+n_res_blocks else residual_out_channels residual_layer = ResidualBlock(residual_in_channels,residual_out_channels) self.residual_layers.append(residual_layer) self.up_sampling_layers = nn.ModuleList() for i in range(n_upsampling_blocks): residual_in_channels = residual_in_channels//2 if i!=0 else residual_in_channels residual_out_channels = residual_out_channels//2 if i!=n_upsampling_blocks-1 else out_channels up_sample_layer = UpsampleLayer(residual_in_channels,residual_out_channels,growth_channels) self.up_sampling_layers.append(up_sample_layer) self.final_conv_1x1=nn.ConvTranspose1d(residual_out_channels,out_channels,kernel_size=1,stride=1,padding=0,bias=False) for m in self.modules(): if isinstance(m,(nn.ConvTranspose1d)): m.weight.data.normal_(0,std=.02/np.sqrt(m.kernel_size[0]*m.kernel_size[1]*m.in_channels)) def forward(self,x): for downsample_layer in self.downsampling_layers: x=downsample_layer(x) for res_block_idx in range(len(self.residual_layers)//(self.n_downsampling_blocks+self.n_upsampling_blocks)): for res_block_idy in range(self.n_downsampling_blocks+self.n_upsampling_blocks): res_block_idz=res_block_idx*self.n_downsampling_blocks+self.n_upsampling_blocks+res_block_idy x=self.residual_layers[res_block_idz](x) for upsample_layer_idx in range(self.n_upsampling_blocks-1,-1,-1): x=self.up_sampling_layers[upsample_layer_idx](x) x=self.final_conv_1x1(x) return x class DownsampleLayer(nn.Module): def __init__(self,in_features,out_features,growth_rate,batch_norm=True): super(DownsampleLayer,self).__init__() self.in_features=in_features self.out_features=out_features self.growth_rate=growth_rate self.conv_7x7_stride_4=nn.ConvTranspose1d(in_features,in_features,kernel_size=7,stride=4,padding=3,bias=False) if batch_norm: self.bn_7x7_stride_4=nn.BatchNorm1d(in_features,momentum=.05) if batch_norm: self.bn_7x7_stride_4.weight.data.fill_(self.lr_mult*10.) if batch_norm: self.bn