scglue.models.glue.GLUETrainer

class scglue.models.glue.GLUETrainer(net, lam_data=None, lam_kl=None, lam_graph=None, lam_align=None, dsc_steps=None, modality_weight=None, optim=None, lr=None, **kwargs)

Bases: Trainer

Trainer for GLUE

Parameters:

• net (scglue.models.glue.GLUE) – GLUE network to be trained

• lam_data (typing.Optional[float]) – Data weight

• lam_graph (typing.Optional[float]) – Graph weight

• lam_align (typing.Optional[float]) – Adversarial alignment weight

• dsc_steps (typing.Optional[int]) – Number of discriminator steps per encoder-decoder step

• modality_weight (typing.Optional[typing.Mapping[str, float]]) – Relative modality weight (indexed by modality name)

• optim (typing.Optional[str]) – Optimizer

• lr (typing.Optional[float]) – Learning rate

• **kwargs – Additional keyword arguments are passed to the optimizer constructor

Methods

compute_losses	Compute loss functions
fit	Fit network
format_data	Format data tensors :rtype: typing.Tuple[typing.Mapping[str, torch.Tensor], typing.Mapping[str, torch.Tensor], torch.Tensor, torch.Tensor, torch.Tensor]
get_losses	Get loss values for given data
load_state_dict	Load state from a state dict
state_dict	State dict
train_step	A single training step
val_step	A single validation step

Attributes

logger