limap.features.models package

Submodules

limap.features.models.base_model module

Base class for models. See mnist_net.py for an example of model.

class limap.features.models.base_model.BaseModel(conf={})

Bases: Module

What the child model is expect to declare:: default_conf: dictionary of the default configuration of the model. It overwrites base_default_conf in BaseModel, and it is overwritten by the user-provided configuration passed to __init__. Configurations can be nested. required_data_keys: list of expected keys in the input data dictionary. strict_conf (optional): boolean. If false, BaseModel does not raise an error when the user provides an unknown configuration entry. _init(self, conf): initialization method, where conf is the final configuration object (also accessible with self.conf). Accessing unkown configuration entries will raise an error. _forward(self, data): method that returns a dictionary of batched prediction tensors based on a dictionary of batched input data tensors. loss(self, pred, data): method that returns a dictionary of losses, computed from model predictions and input data. Each loss is a batch of scalars, i.e. a torch.Tensor of shape (B,). The total loss to be optimized has the key ‘total’. metrics(self, pred, data): method that returns a dictionary of metrics, each as a batch of scalars.

base_default_conf = {'freeze_batch_normalization': False, 'name': None, 'trainable': False}

default_conf = {}

forward(data): Check the data and call the _forward method of the child model.

abstract loss(pred, data): To be implemented by the child class.

abstract metrics(pred, data): To be implemented by the child class.

required_data_keys = []

strict_conf = True

train(mode=True)

Sets the module in training mode.

This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g. Dropout, BatchNorm, etc.

Parameters:: mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.
Returns:: self
Return type:: Module

training: bool

limap.features.models.s2dnet module

class limap.features.models.s2dnet.AdapLayers(hypercolumn_layers: List[str], output_dim: int = 128)

Bases: Module

Small adaptation layers.

forward(features: List[tensor]): Apply adaptation layers.

training: bool

class limap.features.models.s2dnet.S2DNet(conf={})

Bases: BaseModel

default_conf = {'checkpointing': None, 'hypercolumn_layers': ['conv1_2'], 'output_dim': 128, 'pretrained': 's2dnet'}

download_s2dnet_model(path)

loss(pred, data): To be implemented by the child class.

mean = [0.485, 0.456, 0.406]

metrics(pred, data): To be implemented by the child class.

std = [0.229, 0.224, 0.225]

training: bool

limap.features.models.s2dnet.print_gpu_memory()

limap.features.models.vggnet module

class limap.features.models.vggnet.VGGNet(conf={})