qualia_core.evaluation.host package

Submodules

• qualia_core.evaluation.host.Keras module
  • benchmark()
  • Keras
    • Keras.evaluate()
• qualia_core.evaluation.host.Qualia module
  • Result
    • Result.metrics
    • Result.time
  • Qualia
    • Qualia.evaluate()
• qualia_core.evaluation.host.STM32CubeAI module
  • STM32CubeAI
• qualia_core.evaluation.host.TFLite module
  • TFLite
    • TFLite.evaluate()

Module contents

class qualia_core.evaluation.host.Keras(batch_size, repeat, use_gpu, *args, **kwargs)

Bases: object

evaluate(model_kind, dataset, target: str, tag: str, limit: int = None)

class qualia_core.evaluation.host.Qualia(chunks: int | None = None, max_workers: int | None = None, save_preds: bool = False)

Bases: Evaluator

Custom evaluation loop for Qualia host implementations like Qualia-CodeGen Linux example.

Parameters:

save_preds – Save predictions to CSV file in logs/evaluate/<target>

evaluate(framework: LearningFramework[Any], model_kind: str, dataset: RawDataModel, target: str, tag: str, limit: int | None = None, dataaugmentations: list[DataAugmentation] | None = None) → Stats

class qualia_core.evaluation.host.STM32CubeAI(*args, **kwargs)

Bases: STM32CubeAI

class qualia_core.evaluation.host.TFLite(*args, **kwargs)

Bases: object

evaluate(framework, model_kind, dataset, target: str, tag: str, limit: int = None, dataaugmentations=[])