processor package

Submodules

processor.processor module

processor.processor.do_inference(cfg, model, val_loader, num_query)
processor.processor.do_train(cfg, model, center_criterion, train_loader, val_loader, optimizer, optimizer_center, scheduler, loss_fn, num_query)
Args:

training function inputs

cfg: configuration file, passed from /config/configs.py or /default.py

model: initialized deep model

center_criterion: could be enabled if using center loss, implemented for further updating

train_loader: training dataloader

val_loader: validation dataloader

optimizer: SGD or ADAM optimizer

optimizer_center: SGD or ADAM optimizer for center loss

scheduler: updating scheduler

loss_fn: loss function for learning

num_query: number of learning samples

Returns:

processor.processor_bk module

processor.processor_bk.do_inference(cfg, model, val_loader, num_query)
processor.processor_bk.do_train(cfg, model, center_criterion, train_loader, val_loader, optimizer, optimizer_center, scheduler, loss_fn, num_query)

Module contents