🔎 Approach Summary Framework: MMDetection Backbone: Swin Transformer (Large) with Co-DETR head Training Time Constraint: 9 hours Key Insight: A smaller, hi... See more

ApproachOur solution is a very standard object detection workflow. We experimented with a new SOTA model that was launched whilst the competition was going on.... See more

Team: Brain BlendBy @koleshjr & @Sodiq_Babawale_The Amini Cocoa Contamination Challenge tasked participants with developing machine learning models capable... See more

🍫 Introduction to Our YOLO WAY When it comes to cocoa farming, pests and diseases don’t play fair. But neither do we — especially when we’re armed with YOLOv11... See more

Our solution (Team Neural Beans with @100i and me) is conceptually simple - just a single object detection model. We fine-tuned a DINO model with a Swin Transfo... See more

Hello, It has been 9 days since this competition ended and I am curious about the best solutions but I see that it is still under code review so that might be t... See more

I'm very much shocked, even if it is stated that the "Reveal" comes before the "Close" which in itself is very weird, it still very much apears to be one big bu... See more

In the timeline it says that the reveal date is before the closing date! which date is right?

I and my team mate still have pending submissions which refuse to score. @Zindi @Amy_Bray Any way to resolve this?

It's seems Dataset is not properly labelled as i have trained for more than 200 epochs in general there should be change in the precsion and recall but the prec... See more

How are you doing Zindians , I am struggling to install mmcv in my google colab environment, its getting stuck at building wheels , i waited for an hour its sti... See more

Cannot break the 0.76 with yolo11 and various pre-/post-processing. what is the secret? Cannot get to reliably train any mmdet model

Hello everyone! what is your CV-LB scores so far? I have a big gap between them

Per @stefan027 discussion here: https://zindi.africa/competitions/amini-cocoa-contamination-challenge/discussions/26327Take a quick look at this:Please i am hum... See more

[EDIT TO ORIGINAL POST] See this explanation by @Muhamed_Tuo below.---ORIGINAL POST:The labels (bounding boxes) for this challenge are found in two places: As .... See more

Hey Everyone,I would like to know which vision models other contestants have been using and have you been finding any success with them?I have used Yolo and Fas... See more

I tried downloading the dataset, but it's taking some time. I also tried to find a way to download it using Python, and I came across https://api.zindi.africa.... See more

Hey Everyone,When it comes to the submission csv file, are there any other requirements that need to be taken into account such as rounding of the conf value or... See more

Hello. I was wondering whether the generalization task simply refers to avoiding overfitting, or if the model is somehow supposed to classify unrecognized disea... See more

what is the meaning of this sentence:"creating models that can a) generalise well, even when encountering new diseases not seen in the training set," can we en... See more

The total price is 7000 USD but the repartition on the competition info: 1st place: $1 250 USD2nd place: $750 USD3rd place: $500 USD

Can someone be kind enough to suggest a successful strategy for dealing with the class imbalance? I have tried various strategies, upsampling, downsampling, aug... See more

Is there a big brother with an African identity, I invite you to form a team, I am not African

It seems classes and class ids are not matched consistently, see below e.g.

I would appreciate any kind of help please. what do they mean by "a) generalise well, even when encountering new diseases not seen in the training set" does mea... See more