Embeddings param uses hard coded 576 embedding dim and not dim passed

[lanalex]

When passing embeddings=embeddings_matrix (shape = (N, embedding_dim) np array of type float 32 (also ta simple doc example for using this case would be much appreciated :) ). In fastdup.run, the code throws an error that the embedding dim is different form 576 . Because of this The expected behaviour that the embedding dim will be automatically deduced.

class FastdupController:
    @v1_sentry_handler
    def __init__(self, work_dir: Union[str, Path], input_dir: Union[str, Path] = None):
        """
        This class serves as a proxy for fastdup basic usage,
        the class wraps fastdup-run call provides quick access to
        fastdup files such as: similarity,  csv outlier csv, etc...

        Moreover, the class provides several extra features:
            - Ability to run connected component analysis on splits without calling fastdup run again
            - Ability to add annotation file and quickly merge it to any of fastdup inputs
        Currently the class support running fastdup on images and object
        :param work_dir: target output dir or existing output dir
        :param input_dir: (Optional) path to data dir
        """
        # check if fastdup was already applied
        self._fastdup_applied = is_fastdup_dir(work_dir)
        self._work_dir = Path(work_dir)
        self._input_dir = input_dir if input_dir is None else get_input_dir(input_dir)

        # set default arguments
        self._df_annot = None
        self._run_mode = FD.MODE_DEFAULT
        self._embeddings_dim = 576
        self._max_fd_id = 0

[dbickson]

Thanks @lanalex for pointing out this bug and unclear documentation. It will be fixed in 0.917 to be released later today.

[lanalex]

@dbickson Just to clarify the error is not only the dimension. It has some issues that the dataframe input structure. if you can also test it with passing a yolo compatible dataframe (a dataframe with img_filename, bbox_h, bbox_w etc). I had to add an index column (explictly) and modify the code a bit to get it to work. Otherwise it falls on an assert ``` assert 'index' in df_mapping.columns and 'filename' in df_mapping.columns