dev(narugo): regenerate taggers benchmark

import random

from benchmark import BaseBenchmark, create_plot_cli

from imgutils.tagging import get_deepdanbooru_tags, get_wd14_tags, get_mldanbooru_tags, get_deepgelbooru_tags

from imgutils.tagging import get_deepdanbooru_tags, get_wd14_tags, get_mldanbooru_tags, get_deepgelbooru_tags, \

    get_camie_tags

class DeepdanbooruBenchmark(BaseBenchmark):

        _ = get_mldanbooru_tags(image_file)

class CamieBenchmark(BaseBenchmark):

    def __init__(self, model_name):

        BaseBenchmark.__init__(self)

        self.model_name = model_name

    def load(self):

        from imgutils.tagging.camie import _get_camie_model, _get_camie_labels, _get_camie_threshold, \

            _get_camie_preprocessor

        _ = _get_camie_model(self.model_name)

        _ = _get_camie_labels(self.model_name)

        _ = _get_camie_threshold(self.model_name)

        _ = _get_camie_preprocessor(self.model_name)

    def unload(self):

        from imgutils.tagging.camie import _get_camie_model, _get_camie_labels, _get_camie_threshold, \

            _get_camie_preprocessor

        _get_camie_model.cache_clear()

        _get_camie_labels.cache_clear()

        _get_camie_threshold.cache_clear()

        _get_camie_preprocessor.cache_clear()

    def run(self):

        image_file = random.choice(self.all_images)

        _ = get_camie_tags(image_file, model_name=self.model_name)

if __name__ == '__main__':

    create_plot_cli(

        [

            ('wd-vit-large-tagger-v3', Wd14Benchmark("ViT_Large")),

            ('wd-eva02-large-tagger-v3', Wd14Benchmark("EVA02_Large")),

            ('mldanbooru', MLDanbooruBenchmark()),

            ('camie-initial', CamieBenchmark('initial')),

            ('camie-refined', CamieBenchmark('refined')),

        ],

        title='Benchmark for Tagging Models',

        run_times=10,