logger(epoch=0, loss=0, source_train_acc=None, source_val_acc=None, target=None, time=None, verbose=0, train=True)

Print formatted training/testing progress information.

Parameters:
  • epoch (int, default: 0 ) –

    Current training epoch. Default: 0

  • loss (float or tuple, default: 0 ) –

    Loss value(s) for current epoch. If tuple, contains inner and outer losses. Default: 0

  • source_train_acc (float, default: None ) –

    Source domain training accuracy. Default: None

  • source_val_acc (float, default: None ) –

    Source domain validation accuracy. Default: None

  • target (Tensor, default: None ) –

    Target domain predictions/labels. Default: None

  • time (float, default: None ) –

    Time taken for current epoch. Default: None

  • verbose (int, default: 0 ) –

    Verbosity level controlling output detail:

    • 0: No output
    • 1: Basic loss information
    • 2: Add accuracy metrics
    • 3: Add detailed metrics (recall, precision, etc.)

    Default: 0

  • train (bool, default: True ) –

    Whether in training or testing mode. Default: True
Notes

Output Levels:

  • Basic Output (verbose=1):

  • Epoch number (training) or "Test" (testing)

  • Loss values (single or inner/outer)

  • Extended Output (verbose=2):

  • Basic output

  • Source domain accuracy
  • Target domain accuracy

  • Timing information

  • Detailed Output (verbose=3):

  • Extended output

  • Recall at k
  • Precision at k
  • Average precision
  • F1 score
  • Contamination metrics

Features:

  • Multi-level verbosity
  • Flexible metric display
  • Progress tracking
  • Performance monitoring

Format:

  • Epoch XXXX: Loss X.XXXX | Source Acc X.XXXX | Target Acc X.XXXX | Metrics ... | Time X.XX