xenonpy.model.nn package

Submodules

xenonpy.model.nn.layer module

class xenonpy.model.nn.layer.Layer1d(n_in, n_out, *, drop_out=0.0, layer_func=functools.partial(<class 'torch.nn.modules.linear.Linear'>, bias=True), act_func=ReLU(), batch_nor=functools.partial(<class 'torch.nn.modules.batchnorm.BatchNorm1d'>, eps=1e-05, momentum=0.1, affine=True))

Bases: Module

Base NN layer. This is a wrap around PyTorch. See here for details: http://pytorch.org/docs/master/nn.html#

Parameters:

• n_in (int) – Size of each input sample.

• n_out (int) – Size of each output sample

• drop_out (float) – Probability of an element to be zeroed. Default: 0.5

• layer_func (func) – Layers come with PyTorch.

• act_func (func) – Activation function.

• batch_nor (func) – Normalization layers

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.

training: bool

xenonpy.model.nn.wrap module

class xenonpy.model.nn.wrap.Init

Bases: object

static uniform(*, scale=0.1)

class xenonpy.model.nn.wrap.L1

Bases: object

static batch_norm(*args, **kwargs)

Wrapper class for torch.nn.BatchNorm1d. http://pytorch.org/docs/0.3.0/optim.html#torch.nn.BatchNorm1d

static conv(*args, **kwargs)

Wrapper class for torch.nn.Conv1d. http://pytorch.org/docs/0.3.0/optim.html#torch.nn.Conv1d

static instance_norm(*args, **kwargs)

Wrapper class for torch.nn.InstanceNorm1d. http://pytorch.org/docs/0.3.0/optim.html#torch.nn.InstanceNorm1d

static linear(*args, **kwargs)

Wrapper class for torch.nn.Linear. http://pytorch.org/docs/0.3.0/optim.html#torch.nn.Linear

class xenonpy.model.nn.wrap.LrScheduler

Bases: object

static exponential_lr(*args, **kwargs)

Wrapper class for torch.optim.lr_scheduler.ExponentialLR. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.lr_scheduler.ExponentialLR

static lambda_lr(*args, **kwargs)

Wrapper class for torch.optim.lr_scheduler.LambdaLR. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.lr_scheduler.LambdaLR

static multi_step_lr(*args, **kwargs)

Wrapper class for torch.optim.lr_scheduler.MultiStepLR. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.lr_scheduler.MultiStepLR

static reduce_lr_on_plateau(*args, **kwargs)

Wrapper class for torch.optim.lr_scheduler.ReduceLROnPlateau. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.lr_scheduler.ReduceLROnPlateau

static step_lr(*args, **kwargs)

Wrapper class for torch.optim.lr_scheduler.StepLR. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.lr_scheduler.StepLR

class xenonpy.model.nn.wrap.Optim

Bases: object

static ada_delta(*args, **kwargs)

Wrapper class for torch.optim.Adadelta. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.Adadelta

static ada_grad(*args, **kwargs)

Wrapper class for torch.optim.Adagrad. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.Adagrad

static ada_max(*args, **kwargs)

Wrapper class for torch.optim.Adamax. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.Adamax

static adam(*args, **kwargs)

Wrapper class for torch.optim.Adam. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.Adam

static asgd(*args, **kwargs)

Wrapper class for torch.optim.ASGD. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.ASGD

static lbfgs(*args, **kwargs)

Wrapper class for torch.optim.LBFGS. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.LBFGS

static r_prop(*args, **kwargs)

Wrapper class for torch.optim.Rprop. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.Rprop

static rms_prop(*args, **kwargs)

Wrapper class for torch.optim.RMSprop. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.RMSprop

static sgd(*args, **kwargs)

Wrapper class for torch.optim.SGD. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.SGD

static sparse_adam(*args, **kwargs)

Wrapper class for torch.optim.SparseAdam. http://pytorch.org/docs/0.3.0/optim.html#torch.optim.SparseAdam

Module contents