scSLAT.model.train.train_GAN

scSLAT.model.train.train_GAN(wdiscriminator, optimizer_d, embds, batch_d_per_iter=5, anchor_scale=0.8)

GAN training strategy

Parameters:

• wdiscriminator (torch.nn.modules.module.Module) – WGAN

• optimizer_d (torch.optim.optimizer.Optimizer) – WGAN parameters

• embds (typing.List[torch.Tensor]) – list of LGCN embd

• batch_d_per_iter (typing.Optional[int]) – WGAN train iter numbers

• anchor_scale (typing.Optional[float]) – ratio of anchor cells

Return type:

torch.Tensor