30a97dc7c5ae39a0e010f8c9d63714ade624f37c,src/pipelines.py,,preprocessing_inference,#Any#Any#,176

Before Change

                                         },
                                cache_dirpath=config.env.cache_dirpath)

        loader = Step(name="loader",
                      transformer=ImageSegmentationLoader(**config.loader),
                      input_data=["specs"],
                      input_steps=[reader_inference],
                      adapter={"X": ([("reader_inference", "X")]),
                               "y": ([("reader_inference", "y")]),
                               "train_mode": ([("specs", "train_mode")]),
                               },
                      cache_dirpath=config.env.cache_dirpath)
    else:
        xy_inference = Step(name="xy_inference",
                            transformer=XYSplit(**config.xy_splitter[model_name]),
                            input_data=["input", "specs"],

After Change

def preprocessing_inference(config, model_name="unet"):
    if config.execution.loader_mode == "crop_and_pad":
        Loader = loaders.ImageSegmentationLoaderCropPad
    elif config.execution.loader_mode == "resize":
        Loader = loaders.ImageSegmentationLoaderResize
    else:
        raise NotImplementedError

    if config.loader.dataset_params.image_source == "memory":
        reader_inference = Step(name="reader_inference",
                                transformer=ImageReader(**config.reader[model_name]),
                                input_data=["input", "specs"],

Instances