1. Liga U19 stats & predictions

Player Spotlight: Rising Stars of Tomorrow

The future of football is bright with young talents emerging every season in the U19 league. Here's a closer look at some promising players who are expected to make waves in top-tier football soon:

• Lukas Hradecky (Goalkeeper): Known for his exceptional reflexes and commanding presence in goal, Lukas has been instrumental in AC Sparta Prague's solid defensive record this season.
• Milan Skoda (Midfielder): Milan's vision on the pitch is unmatched. His ability to control the tempo of the game makes him a crucial asset for FC Viktoria Plzeň's midfield dynamics.
• Ondrej Duda (Forward): With his flair and knack for finding the back of the net, Ondrej is quickly becoming one of the most feared forwards in the league.
• Jakub Jankto (Midfielder): Jakub's versatility allows him to play both as a defensive midfielder and an attacking midfielder. His work rate and tactical awareness are commendable assets for any team he plays for.
• Marek Rodak (Goalkeeper): Known for his agility and shot-stopping abilities, Marek has been keeping clean sheets consistently this season while playing for FC Slovan Liberec.Siegmar/PiNet<|file_sep|>/PiNet.py import torch.nn as nn import torch from torch.autograd import Variable import math from torch.nn import init import numpy as np # Bottleneck residual unit class ResidualBlock(nn.Module): def __init__(self,in_planes,out_planes,stride=1): super(ResidualBlock,self).__init__() self.conv1 = nn.Conv3d(in_planes,out_planes,kernel_size=3,stride=stride,padding=1,bias=False) self.bn1 = nn.BatchNorm3d(out_planes) self.conv2 = nn.Conv3d(out_planes,out_planes,kernel_size=3,stride=1,padding=1,bias=False) self.bn2 = nn.BatchNorm3d(out_planes) self.shortcut = nn.Sequential() if stride !=1 or in_planes != out_planes: self.shortcut = nn.Sequential( nn.Conv3d(in_planes,out_planes,kernel_size=1,stride=stride,bias=False), nn.BatchNorm3d(out_planes) ) def forward(self,x): out = torch.relu(self.bn1(self.conv1(x))) out = self.bn2(self.conv2(out)) out += self.shortcut(x) out = torch.relu(out) return out # Net structure class PiNet(nn.Module): def __init__(self,num_classes=11,num_filters=[16]*4,num_blocks=[3]*4): super(PiNet,self).__init__() self.num_classes = num_classes self.num_filters = num_filters self.num_blocks = num_blocks self.in_planes = self.num_filters[0] self.conv1 = nn.Conv3d(1,self.num_filters[0],kernel_size=7,stride=(1,4,4),padding=(3,12,12),bias=False) self.layers = self._make_layer(self.num_filters[0],self.num_blocks[0],stride=1) for i in range(1,len(self.num_blocks)): self.layers.add_module(str(i+1),self._make_layer(self.num_filters[i],self.num_blocks[i],stride=2)) self.avgpool = nn.AvgPool3d((14,7,7)) self.fc = nn.Linear(self.num_filters[-1],self.num_classes) def _make_layer(self,in_planes,num_blocks,stride): strides = [stride] + [1]*(num_blocks-1) layers = nn.Sequential() for stride in strides: layers.add_module(str(stride),ResidualBlock(in_planes,in_planes*stride,stride)) in_planes *= stride return layers def forward(self,x): x = torch.relu(self.conv1(x)) x = self.layers(x) # x = self.avgpool(x) # # x = x.view(x.size(0),-1) # # x = self.fc(x) return x class PiNetMulti(nn.Module): def __init__(self,num_classes=[11]*5,num_filters=[16]*4,num_blocks=[3]*4): super(PiNetMulti,self).__init__() self.num_classes = num_classes self.num_filters = num_filters self.num_blocks = num_blocks self.in_planes = self.num_filters[0] self.conv1_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0 = nn.Conv3d(12,self.num_filters[0],kernel_size=7,stride=(1,4,4),padding=(3,12,12),bias=False) # change here if using other input channels self.layers = self._make_layer(self.num_filters[0],self.num_blocks[0],stride=1) for i in range(1,len(self.num_blocks)): self.layers.add_module(str(i+1),self._make_layer(self.num_filters[i],self.num_blocks[i],stride=2)) self.avgpool = nn.AvgPool3d((14,7,7)) # This will be changed according to different output classes # The input channel will be equal to number of output classes from previous network # The output channel will be equal to number of output classes from current network # Convolution layer between PiNet-4 output class and PiNet-5 input class # Example: # If PiNet-4 has three output classes then number of input channels will be three. # If PiNet-5 has five output classes then number of output channels will be five. # If PiNet-4 has two output classes then number of input channels will be two. # If PiNet-5 has three output classes then number of input channels will be three. self.conv5to6layerinputchannellengthlist=[] conv5to6layerinputchannellengthlist=[] conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) conv5to6layerinputchannellengthlist.append(num_classes[4]) #print(conv5to6layerinputchannellengthlist) self.conv5to6layeroutputchannellengthlist=[] conv5to6layeroutputchannellengthlist=[] conv5to6layeroutputchannellengthlist.append(num_classes[5]) conv5to6layeroutputchannellengthlist.append(num_classes[5]) conv5to6layeroutputchannellengthlist.append(num_classes[5]) conv5to6layeroutputchannellengthlist.append(num_classes[5]) conv5to6layeroutputchannellengthlist.append(num_classes[5]) conv5to6layeroutputchannellengthlist.append(num_classes[5]) conv5to6layeroutputchannellengthlist.append(num_classes[5]) print(conv5to6layeroutputchannellengthlist) sumofconvchannelsbeforeconv=[] sumofconvchannelsbeforeconv=[sum(conv5to6layerinputchannellengthlist[:i]) for i in range(len(conv5to6layerinputchannellengthlist)+1)] print(sumofconvchannelsbeforeconv) print(len(sumofconvchannelsbeforeconv)) self.conv56a=self._make_layer(sumofconvchannelsbeforeconv[-1],num_blocks[str(56)],stride=56) sumofconvchannelsbeforeconv=sumofconvchannelsbeforeconv[:-1] print(sumofconvchannelsbeforeconv) print(len(sumofconvchannelsbeforeconv)) # This will be changed according to different output classes # The input channel will be equal to number of output classes from previous network # The output channel will be equal to number of output classes from current network # Convolution layer between PiNet-3 output class and PiNet-4 input class # Example: # If PiNet-3 has two output classes then number of input channels will be two. # If PiNet-4 has three output classes then number of input channels will be three. self.conv34a=self._make_layer(sumofconvchannelsbeforeconv[-1],num_blocks[str(34)],stride=34) sumofconvchannelsbeforeconv=sumofconvchannelsbeforeconv[:-1] print(sumofconvchannelsbeforeconv) print(len(sumofconvchannelsbeforeconv))