Update in Evaluator

F1 CAR DETECTION-26055d
#1

Hi all!

We have now updated the evaluator to take the right format from the submission file (x_min, y_min, x_max, y_max) and have completed re-evaluations.

For those who are seeing lower score as compared to previous score, please update the submission file from the format (x_min, x_max, y_min, y_max ) to (x_min, y_min, x_max, y_max).

Here the code to convert the submission into the right format

# Importing Libraries
import pandas as pd
from ast import literal_eval

# Reading the submission file
submission = pd.read_csv("submission.csv")


# Going through each row
for index, row in submission.iterrows():

    # Convert the bounding boxes from string to python list
    bboxes = literal_eval(row["bboxes"])[0]

    # Convert the bounding boxes to the right format
    new_bboxes = [[bboxes[0], bboxes[2], bboxes[1], bboxes[3], bboxes[4]]]

    # Making changes to the row
    submission["bboxes"][index] = new_bboxes

# Saving the changes into a new submission file
submission.to_csv("new_submisson.csv", index=False)

Sorry for the inconvenience caused.

Regards,
Shubhamai

1 Like