hyrax.models.simclr

Classes

NTXentLoss	Normalized Temperature-scaled Cross Entropy Loss. Based on Chen, 2020
PositiveRescale	Transformation Class specifically for ColorJitter to prevent wrong domain during the augmentation
SimCLR	SimCLR model. Implementation based on Chen, 2020

Module Contents

class NTXentLoss(temperature=0.1)

Bases: torch.nn.Module

Normalized Temperature-scaled Cross Entropy Loss. Based on Chen, 2020

Initialize internal Module state, shared by both nn.Module and ScriptModule.

temperature = 0.1

criterion

forward(z_i, z_j)

Forward function of NTXentLoss. Based on Chen, 2020. Loss is calculated from representations from two augmented views of the same batch.

class PositiveRescale(transform)

Transformation Class specifically for ColorJitter to prevent wrong domain during the augmentation

transform

__call__(x)

class SimCLR(config, shape)

Bases: torch.nn.Module

SimCLR model. Implementation based on Chen, 2020

Initialize internal Module state, shared by both nn.Module and ScriptModule.

config

shape

backbone

projection_head

criterion

forward(x)

train_step(x)