Commit 7704e624 authored by narugo1992's avatar narugo1992
Browse files

dev(narugo): update plots

parent e317a073

zoo/ccip/plot.py

+1 −1
Original line number Diff line number Diff line
@@ -25,7 +25,7 @@ def _pos_neg_to_true_score(pos, neg): 




def plt_confusion_matrix(ax, y_true, y_pred, title: str = 'Confusion Matrix',

                         normalize: Literal['true', 'pred', None] = None, cmap=None):

                         normalize: Literal['true', 'pred', None] = 'true', cmap=None):

    cm = confusion_matrix(y_true, y_pred, normalize=normalize)

    disp = ConfusionMatrixDisplay(

        confusion_matrix=cm,

zoo/ccip/publish.py

+21 −11
Original line number Diff line number Diff line
@@ -17,7 +17,7 @@ from huggingface_hub import hf_hub_download, HfApi, CommitOperationAdd 
from lighttuner.hpo import hpo, R, uniform, randint

from natsort import natsorted

from sklearn.cluster import OPTICS, DBSCAN

from sklearn.metrics import adjusted_rand_score

from sklearn.metrics import adjusted_rand_score, precision_score, recall_score

from torchvision import transforms

from tqdm.auto import tqdm
@@ -148,12 +148,26 @@ def create_plots(dist, cmatrix): 
    y_true = cmatrix.reshape(-1).type(torch.int).numpy()

    y_pred = (dist <= threshold).reshape(-1).type(torch.int).numpy()

    accuracy = (y_true == y_pred).sum() / y_true.shape[0]

    logging.info(f'Threshold: {threshold:.4f}, f1 score: {f1_score:.4f}, accuracy: {accuracy * 100.0:.2f}%')

    precision = precision_score(y_true, y_pred, pos_label=1)

    recall = recall_score(y_true, y_pred, pos_label=1)

    metrics = {

        'threshold': threshold,

        'f1_score': f1_score,

        'accuracy': accuracy,

        'precision': precision,

        'recall': recall,

    }

    logging.info(f'Threshold: {threshold:.4f}, f1 score: {f1_score:.4f}, '

                 f'precision: {precision * 100.0:.2f}%, recall: {recall * 100.0:.2f}%')



    logging.info('Creating confusion matrix ...')

    plots['confusion_matrix'] = plt_export(

    plots['confusion_matrix_true'] = plt_export(

        plt_confusion_matrix, y_true, y_pred,

        normalize='true', title=f'Confusion Matrix (True)'

    )

    plots['confusion_matrix_pred'] = plt_export(

        plt_confusion_matrix, y_true, y_pred,

        title=f'Confusion Matrix\nAccuracy: {accuracy * 100.0:.2f}%'

        normalize='pred', title=f'Confusion Matrix (Predict)'

    )



    logging.info('Creating ROC Curve ...')
@@ -171,7 +185,7 @@ def create_plots(dist, cmatrix): 
    logging.info('Creating F1 Curve ...')

    plots['f1'] = plt_export(plt_f1_curve, pos, neg)



    return (threshold, f1_score, accuracy), plots

    return threshold, metrics, plots





def export_model_to_dir(file_in_repo: str, output_dir: str, repository: str = 'deepghs/ccip',
@@ -188,15 +202,11 @@ def export_model_to_dir(file_in_repo: str, output_dir: str, repository: str = 'd 


    scale = get_scale_for_model(model)

    dist, cids, cmatrix = _get_dist_matrix(model, scale)

    (threshold, f1_score, accuracy), plots = create_plots(dist, cmatrix)

    threshold, metrics, plots = create_plots(dist, cmatrix)

    metrics_file = os.path.join(output_dir, 'metrics.json')

    logging.info(f'Creating metric file {metrics_file!r} ...')

    with open(metrics_file, 'w') as f:

        json.dump({

            'threshold': threshold,

            'f1_score': f1_score,

            'accuracy': accuracy,

        }, fp=f, indent=4, sort_keys=True, ensure_ascii=False)

        json.dump(metrics, fp=f, indent=4, sort_keys=True, ensure_ascii=False)



    clustering_file = os.path.join(output_dir, 'cluster.json')

    logging.info(f'Creating clustering measurement {clustering_file!r} ...')