pyanomaly.networks.auxiliary.flownet2 package

Subpackages

• pyanomaly.networks.auxiliary.flownet2.channelnorm_package package
  • Submodules
  • pyanomaly.networks.auxiliary.flownet2.channelnorm_package.channelnorm module
  • pyanomaly.networks.auxiliary.flownet2.channelnorm_package.setup module
  • Module contents
• pyanomaly.networks.auxiliary.flownet2.correlation_package package
  • Submodules
  • pyanomaly.networks.auxiliary.flownet2.correlation_package.correlation module
  • pyanomaly.networks.auxiliary.flownet2.correlation_package.setup module
  • Module contents
• pyanomaly.networks.auxiliary.flownet2.resample2d_package package
  • Submodules
  • pyanomaly.networks.auxiliary.flownet2.resample2d_package.resample2d module
  • pyanomaly.networks.auxiliary.flownet2.resample2d_package.setup module
  • Module contents

Submodules

pyanomaly.networks.auxiliary.flownet2.FlowNetC module

class pyanomaly.networks.auxiliary.flownet2.FlowNetC.FlowNetC(cfg, batchNorm=True, div_flow=20)

Bases: torch.nn.modules.module.Module

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

pyanomaly.networks.auxiliary.flownet2.FlowNetFusion module

class pyanomaly.networks.auxiliary.flownet2.FlowNetFusion.FlowNetFusion(cfg, batchNorm=True)

Bases: torch.nn.modules.module.Module

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

pyanomaly.networks.auxiliary.flownet2.FlowNetS module

Portions of this code copyright 2017, Clement Pinard

class pyanomaly.networks.auxiliary.flownet2.FlowNetS.FlowNetS(cfg, input_channels=12, batchNorm=True)

Bases: torch.nn.modules.module.Module

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

pyanomaly.networks.auxiliary.flownet2.FlowNetSD module

class pyanomaly.networks.auxiliary.flownet2.FlowNetSD.FlowNetSD(cfg, batchNorm=True)

Bases: torch.nn.modules.module.Module

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

pyanomaly.networks.auxiliary.flownet2.models module

class pyanomaly.networks.auxiliary.flownet2.models.FlowNet2(cfg, batchNorm=False, div_flow=20.0)

Bases: torch.nn.modules.module.Module

forward(inputs)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

init_deconv_bilinear(weight)

class pyanomaly.networks.auxiliary.flownet2.models.FlowNet2C(args, batchNorm=False, div_flow=20)

Bases: pyanomaly.networks.auxiliary.flownet2.FlowNetC.FlowNetC

forward(inputs)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class pyanomaly.networks.auxiliary.flownet2.models.FlowNet2CS(args, batchNorm=False, div_flow=20.0)

Bases: torch.nn.modules.module.Module

forward(inputs)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class pyanomaly.networks.auxiliary.flownet2.models.FlowNet2CSS(args, batchNorm=False, div_flow=20.0)

Bases: torch.nn.modules.module.Module

forward(inputs)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class pyanomaly.networks.auxiliary.flownet2.models.FlowNet2S(args, batchNorm=False, div_flow=20)

Bases: pyanomaly.networks.auxiliary.flownet2.FlowNetS.FlowNetS

forward(inputs)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class pyanomaly.networks.auxiliary.flownet2.models.FlowNet2SD(args, batchNorm=False, div_flow=20)

Bases: pyanomaly.networks.auxiliary.flownet2.FlowNetSD.FlowNetSD

forward(inputs)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

pyanomaly.networks.auxiliary.flownet2.submodules module

pyanomaly.networks.auxiliary.flownet2.submodules.conv(batchNorm, in_planes, out_planes, kernel_size=3, stride=1)

pyanomaly.networks.auxiliary.flownet2.submodules.deconv(in_planes, out_planes)

pyanomaly.networks.auxiliary.flownet2.submodules.i_conv(batchNorm, in_planes, out_planes, kernel_size=3, stride=1, bias=True)

pyanomaly.networks.auxiliary.flownet2.submodules.init_deconv_bilinear(weight)

pyanomaly.networks.auxiliary.flownet2.submodules.predict_flow(in_planes)

pyanomaly.networks.auxiliary.flownet2.submodules.save_grad(grads, name)

class pyanomaly.networks.auxiliary.flownet2.submodules.tofp16

Bases: torch.nn.modules.module.Module

forward(input)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class pyanomaly.networks.auxiliary.flownet2.submodules.tofp32

Bases: torch.nn.modules.module.Module

forward(input)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

Module contents