Amini Cocoa Contamination Challenge
Helping Ghana
$7 000 USD
Completed (11 months ago)
Computer Vision
Object Detection
928 joined
255 active
Start
Feb 14, 25
Close
May 11, 25
Reveal
May 12, 25
'%3e%3cpath fill='%23001489' d='M0 0v6h9V0z'/%3e%3cpath fill='%23e03c31' d='M0 0v3h9V0z'/%3e%3cg stroke='%23fff' stroke-width='2'%3e%3cpath id='d' d='m0 0 4.5 3L0 6m4.5-3H9'/%3e%3cuse xlink:href='%23b' stroke='%23ffb81c' clip-path='url(%23c)'/%3e%3c/g%3e%3cuse xlink:href='%23d' fill='none' stroke='%23007749' stroke-width='1.2'/%3e%3c/g%3e%3c/svg%3e){kind=small}
Classes not unique
Data · 6 Mar 2025, 06:58 · 7
It seems classes and class ids are not matched consistently, see below e.g.
I believe the correct IDs are in the labels folder
Yeah, I guess so, but ... those are all numeric.
So is a label of 2 or 1 healthy? Is cssvd 0 or 1?
Thank you for pointing this error out. Here is a quick fix that you can include in your notebook:
# strip any spacing from the class item and make sure it is a string
train['class'] = train['class'].str.strip()
# The correct mapping from class to class_id
class_map = {cls: i for i, cls in enumerate(sorted(train['class'].unique().tolist()))}
# This will give you
{'anthracnose': 0, 'cssvd': 1, 'healthy': 2}
# Map it
train['class_id'] = train['class'].map(class_map)
# Check
train[['class', 'class_id']].value_counts()
class class_id count
healthy 2 4280
cssvd 1 3241
anthracnose 0 2271
Great, thanks for clarifying
Thank you! Some images in the Train / csv sometimes no longer have annotations in the associated label file. We can clearly see in the train that an image is associated with cssvd and healthy but in the associated label file we no longer see cssvd. Is this an anomaly or a misunderstanding on my part?