qualia_core.experimenttracking.pytorch package

Submodules

• qualia_core.experimenttracking.pytorch.ClearML module
  • ClearML
    • ClearML.start()
    • ClearML.stop()
    • ClearML.initializer()
    • ClearML.import_and_clear_all_offline_sessions()
    • ClearML.logger
• qualia_core.experimenttracking.pytorch.ExperimentTrackingPyTorch module
  • ExperimentTrackingPyTorch
    • ExperimentTrackingPyTorch.logger
• qualia_core.experimenttracking.pytorch.Neptune module
  • Neptune
    • Neptune.start()
    • Neptune.stop()
    • Neptune.logger

Module contents

class qualia_core.experimenttracking.pytorch.ClearML(project_name: str, task_name: str, sources_path: str | None = None, ignores: list[str] | None = None, offline_mode: bool = False)

Bases: ExperimentTracking

classmethod import_and_clear_all_offline_sessions() → None

classmethod initializer() → None

Connect to task in PyTorch Lightning Trainer subprocess with e.g. ddp_spawn.

property logger: None

start(name: str | None = None) → None

stop() → None

class qualia_core.experimenttracking.pytorch.Neptune(project_name: str, config_file: str = 'conf/neptune.toml')

Bases: NeptuneBase, ExperimentTrackingPyTorch

property logger: NeptuneLogger

start(name: str | None = None) → None

stop() → None

class qualia_core.experimenttracking.pytorch.QualiaDatabase(db_path: str | Path | None = None)

Bases: ExperimentTracking

classmethod cli() → None

get_plugins() → list[dict[str, Any]]

handle_command(command: str, *args: str) → str

log_quantization(model_id: int, bits: int, epochs: int) → None

log_stats(model_name: str, model_hash: str, stats: Stats) → None

log_trainresult(trainresult: TrainResult) → int | None

property logger: None

classmethod print_cli_help() → str

start(name: str | None = None) → None

stop() → None