{"id":297796,"date":"2020-01-28T09:00:26","date_gmt":"2020-01-28T09:00:26","guid":{"rendered":"http:\/\/savepearlharbor.com\/?p=297796"},"modified":"-0001-11-30T00:00:00","modified_gmt":"-0001-11-29T21:00:00","slug":"","status":"publish","type":"post","link":"https:\/\/savepearlharbor.com\/?p=297796","title":{"rendered":"\u041a\u0430\u043a \u0441\u0434\u0435\u043b\u0430\u0442\u044c \u0431\u043e\u0442\u0430, \u043a\u043e\u0442\u043e\u0440\u044b\u0439 \u043f\u0440\u0435\u0432\u0440\u0430\u0449\u0430\u0435\u0442 \u0444\u043e\u0442\u043e \u0432 \u043a\u043e\u043c\u0438\u043a\u0441. \u0427\u0430\u0441\u0442\u044c \u0442\u0440\u0435\u0442\u044c\u044f. Serverless GPU \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u043c\u043e\u0434\u0435\u043b\u0438"},"content":{"rendered":"\n
\n

-> \u0427\u0430\u0441\u0442\u044c 1<\/a>
-> \u0427\u0430\u0441\u0442\u044c 2<\/a><\/p>\n

<\/p>\n

\u041d\u0443, \u043e\u0442\u0434\u043e\u0445\u043d\u0443\u043b\u0438 \u0438 \u0445\u0432\u0430\u0442\u0438\u0442. \u0421 \u0432\u043e\u0437\u0432\u0440\u0430\u0449\u0435\u043d\u0438\u0435\u043c! <\/p>\n

<\/p>\n

\u0412 \u043f\u0440\u0435\u0434\u044b\u0434\u0443\u0449\u0438\u0445 \u0441\u0435\u0440\u0438\u044f\u0445 \u043c\u044b \u0441 \u0432\u0430\u043c\u0438 \u0441\u043e\u0431\u0440\u0430\u043b\u0438 \u0434\u0430\u043d\u043d\u044b\u0435 \u0438 \u043e\u0431\u0443\u0447\u0438\u043b\u0438 \u0441\u0432\u043e\u044e \u043f\u0435\u0440\u0432\u0443\u044e \u043c\u043e\u0434\u0435\u043b\u044c.
\u0417\u0430\u0442\u0435\u043c, \u0443\u0436\u0430\u0441\u043d\u0443\u0432\u0448\u0438\u0441\u044c \u0440\u0435\u0437\u0443\u043b\u044c\u0442\u0430\u0442\u0430\u043c, \u043e\u0431\u0443\u0447\u0438\u043b\u0438 \u0435\u0449\u0435 \u0441 \u0434\u0435\u0441\u044f\u0442\u043e\u043a.
\u0421\u0430\u043c\u043e\u0435 \u0432\u0440\u0435\u043c\u044f \u043f\u043e\u043a\u0430\u0437\u0430\u0442\u044c \u043d\u0430\u0448\u0435 \u0442\u0432\u043e\u0440\u0435\u043d\u0438\u0435 \u043c\u0438\u0440\u0443!<\/a><\/p>\n

<\/p>\n

\u042d\u043a\u0441\u043f\u043e\u0440\u0442 \u043c\u043e\u0434\u0435\u043b\u0438<\/h2>\n

<\/p>\n

\u0414\u043b\u044f \u043d\u0430\u0447\u0430\u043b\u0430 \u043f\u0435\u0440\u0435\u0441\u043e\u0445\u0440\u0430\u043d\u0438\u043c \u043c\u043e\u0434\u0435\u043b\u044c \u0433\u0435\u043d\u0435\u0440\u0430\u0442\u043e\u0440\u0430 \u0432 \u043f\u043e\u0434\u0445\u043e\u0434\u044f\u0449\u0438\u0439 \u0444\u043e\u0440\u043c\u0430\u0442, \u0447\u0442\u043e\u0431\u044b \u043d\u0430\u043c \u043d\u0435 \u043f\u0440\u0438\u0448\u043b\u043e\u0441\u044c \u0442\u0430\u0449\u0438\u0442\u044c \u0434\u0435\u043a\u043b\u0430\u0440\u0430\u0446\u0438\u0438 \u043a\u043b\u0430\u0441\u0441\u043e\u0432 \u043d\u0430 \u0445\u043e\u0441\u0442\u0438\u043d\u0433.<\/p>\n

<\/p>\n

\u0421\u043e\u0437\u0434\u0430\u0434\u0438\u043c \u043d\u0435\u0431\u043e\u043b\u044c\u0448\u043e\u0439 \u0444\u0430\u0439\u043b\u0438\u043a \u0441 \u0440\u0430\u0441\u0448\u0438\u0440\u0435\u043d\u0438\u0435\u043c *.py \u0438 \u0441\u043a\u043e\u043f\u0438\u0440\u0443\u0435\u043c \u0432 \u043d\u0435\u0433\u043e \u043a\u043e\u0434 \u0438\u0437-\u043f\u043e\u0434 \u0441\u043f\u043e\u0439\u043b\u0435\u0440\u0430 \u043d\u0438\u0436\u0435.
\u041f\u0443\u0441\u0442\u044c \u044d\u0442\u043e \u0431\u0443\u0434\u0435\u0442 jit.py:<\/p>\n

<\/p>\n

\u041a\u043e\u0434, \u043a\u043e\u0442\u043e\u0440\u044b\u0439 \u043d\u0443\u0436\u043d\u043e \u0431\u0435\u0437\u0434\u0443\u043c\u043d\u043e \u0441\u043a\u043e\u043f\u0438\u043f\u0430\u0441\u0442\u0438\u0442\u044c, \u0437\u0430\u043c\u0435\u043d\u0438\u0432 path, output_path \u043d\u0430 \u0441\u0432\u043e\u0438<\/b><\/p>\n
\n
# \u0417\u0430\u043c\u0435\u043d\u044f\u0435\u043c path \u043d\u0430 \u043f\u0443\u0442\u044c \u043a \u043f\u043e\u043d\u0440\u0430\u0432\u0438\u0432\u0448\u0435\u0439\u0441\u044f \u0438\u0442\u0435\u0440\u0430\u0446\u0438\u0438 \u043c\u043e\u0434\u0435\u043b\u0438 # \u041d\u0430\u043c \u043d\u0443\u0436\u0435\u043d \u0444\u0430\u0439\u043b *G_A.pth - \u0433\u0435\u043d\u0435\u0440\u0430\u0442\u043e\u0440 \u0444\u043e\u0442\u043e -> \u043a\u043e\u043c\u0438\u043a\u0441 # output_path - \u0438\u043c\u044f \u0444\u0430\u0439\u043b\u0430 \u0434\u043b\u044f \u044d\u043a\u0441\u043f\u043e\u0440\u0442\u0438\u0440\u0443\u0435\u043c\u043e\u0439 \u043c\u043e\u0434\u0435\u043b\u0438 # \u043c\u043e\u0436\u0435\u0442 \u0431\u044b\u0442\u044c \u043b\u044e\u0431\u044b\u043c \u0441 \u0440\u0430\u0441\u0448\u0438\u0440\u0435\u043d\u0438\u0435\u043c *.jit, \u0433\u043b\u0430\u0432\u043d\u043e\u0435 - \u043d\u0435 \u0437\u0430\u0431\u044b\u0442\u044c, \u043a\u0443\u0434\u0430 \u043c\u044b \u0435\u0433\u043e \u0441\u043e\u0445\u0440\u0430\u043d\u0438\u043b\u0438  path= '\/checkpoints\/resnet9_nowd_nodo_128to400_c8\/60_net_G_A.pth' output_path ='\/checkpoints\/resnet9_nowd_nodo_128to400_c8\/resnet9_nowd_nodo_128to400_c8_160-50-60_1.jit'  import torch from torch import nn  class ResnetGenerator(nn.Module):     """Resnet-based generator that consists of Resnet blocks between a few downsampling\/upsampling operations.      We adapt Torch code and idea from Justin Johnson's neural style transfer project(https:\/\/github.com\/jcjohnson\/fast-neural-style)     """      def __init__(self, input_nc, output_nc, ngf=64, norm_layer=nn.BatchNorm2d, use_dropout=False, n_blocks=6, padding_type='reflect'):         """Construct a Resnet-based generator          Parameters:             input_nc (int)      -- the number of channels in input images             output_nc (int)     -- the number of channels in output images             ngf (int)           -- the number of filters in the last conv layer             norm_layer          -- normalization layer             use_dropout (bool)  -- if use dropout layers             n_blocks (int)      -- the number of ResNet blocks             padding_type (str)  -- the name of padding layer in conv layers: reflect | replicate | zero         """         assert(n_blocks >= 0)         super(ResnetGenerator, self).__init__()         if type(norm_layer) == functools.partial:             use_bias = norm_layer.func == nn.InstanceNorm2d         else:             use_bias = norm_layer == nn.InstanceNorm2d          model = [nn.ReflectionPad2d(3),                  nn.Conv2d(input_nc, ngf, kernel_size=7, padding=0, bias=use_bias),                  norm_layer(ngf),                  nn.ReLU(True)]          n_downsampling = 2         for i in range(n_downsampling):  # add downsampling layers             mult = 2 ** i             model += [nn.Conv2d(ngf * mult, ngf * mult * 2, kernel_size=3, stride=2, padding=1, bias=use_bias),                       norm_layer(ngf * mult * 2),                       nn.ReLU(True)]          mult = 2 ** n_downsampling         for i in range(n_blocks):       # add ResNet blocks              model += [ResnetBlock(ngf * mult, padding_type=padding_type, norm_layer=norm_layer, use_dropout=use_dropout, use_bias=use_bias)]          for i in range(n_downsampling):  # add upsampling layers             mult = 2 ** (n_downsampling - i)             model += [nn.ConvTranspose2d(ngf * mult, int(ngf * mult \/ 2),                                          kernel_size=3, stride=2,                                          padding=1, output_padding=1,                                          bias=use_bias),                       norm_layer(int(ngf * mult \/ 2)),                       nn.ReLU(True)]         model += [nn.ReflectionPad2d(3)]         model += [nn.Conv2d(ngf, output_nc, kernel_size=7, padding=0)]         model += [nn.Tanh()]          self.model = nn.Sequential(*model)      def forward(self, input):         """Standard forward"""         return self.model(input)  class ResnetBlock(nn.Module):     """Define a Resnet block"""      def __init__(self, dim, padding_type, norm_layer, use_dropout, use_bias):         """Initialize the Resnet block          A resnet block is a conv block with skip connections         We construct a conv block with build_conv_block function,         and implement skip connections in <forward> function.         Original Resnet paper: https:\/\/arxiv.org\/pdf\/1512.03385.pdf         """         super(ResnetBlock, self).__init__()         self.conv_block = self.build_conv_block(dim, padding_type, norm_layer, use_dropout, use_bias)      def build_conv_block(self, dim, padding_type, norm_layer, use_dropout, use_bias):         """Construct a convolutional block.          Parameters:             dim (int)           -- the number of channels in the conv layer.             padding_type (str)  -- the name of padding layer: reflect | replicate | zero             norm_layer          -- normalization layer             use_dropout (bool)  -- if use dropout layers.             use_bias (bool)     -- if the conv layer uses bias or not          Returns a conv block (with a conv layer, a normalization layer, and a non-linearity layer (ReLU))         """         conv_block = []         p = 0         if padding_type == 'reflect':             conv_block += [nn.ReflectionPad2d(1)]         elif padding_type == 'replicate':             conv_block += [nn.ReplicationPad2d(1)]         elif padding_type == 'zero':             p = 1         else:             raise NotImplementedError('padding [%s] is not implemented' % padding_type)          conv_block += [nn.Conv2d(dim, dim, kernel_size=3, padding=p, bias=use_bias), norm_layer(dim), nn.ReLU(True)]         if use_dropout:             conv_block += [nn.Dropout(0.5)]          p = 0         if padding_type == 'reflect':             conv_block += [nn.ReflectionPad2d(1)]         elif padding_type == 'replicate':             conv_block += [nn.ReplicationPad2d(1)]         elif padding_type == 'zero':             p = 1         else:             raise NotImplementedError('padding [%s] is not implemented' % padding_type)         conv_block += [nn.Conv2d(dim, dim, kernel_size=3, padding=p, bias=use_bias), norm_layer(dim)]          return nn.Sequential(*conv_block)      def forward(self, x):         """Forward function (with skip connections)"""         out = x + self.conv_block(x)  # add skip connections         return out  import functools  norm_layer = functools.partial(nn.InstanceNorm2d, affine=False, track_running_stats=False)  model = ResnetGenerator(3,3,64, norm_layer=norm_layer, use_dropout=False, n_blocks=9)  model.load_state_dict(torch.load(path)) model.eval() model.cuda() model.half()  trace_input = torch.ones(1,3,256,256).cuda() model = model.eval() jit_model = torch.jit.trace(model.float(), trace_input)  torch.jit.save(jit_model, output_path)<\/code><\/pre>\n<\/div>\n<\/div>\n
  <\/p>\n
\u0417\u0430\u043c\u0435\u043d\u044f\u0435\u043c \u043f\u0435\u0440\u0435\u043c\u0435\u043d\u043d\u044b\u0435 \u043d\u0430 \u0441\u0432\u043e\u0438: <\/p>\n
  <\/p>\n