Lacuna Solar Survey ChallengeCan you identify solar panels and boilers in satellite and drone imagery?1.Dataset1.1.Initial analysis revealed annotation flaws in... See more

I want to express how deeply saddened, disheartened, and concerned I am by recent events related to the competition and how my feedback was handled.As a Black m... See more

First of all, I want to thank the @Zindi team for this competition. However, after the deadline and checking the private leaderboard, I discovered a serious mis... See more

Hello everyone! What approach did you take to solve this problem?Well, I used two models: YOLO for image segmentation and polygon estimation. The idea is to us... See more

I watched a video by @Snow where he explained how to use Gemma3 to count panels. That got me thinking—why not turn it into a full code? If anyone wants to see G... See more

First, I apologize. In a recent post, I mentioned that YOLO could achieve an LB of 0.80, which was based on my past experience and intuition. However, the reali... See more

while we are burning GPU credits and waiting for cv scores, here is one fun video to kill time :Phttps://www.youtube.com/watch?v=haDxH5g8G-M

if you have good rescources train a yolo with image size of 1024 for almost 100 epochs and i can certainly get you under 0.80 and all without usingtabular data... See more

what are your results and thoughts on using that

Why my rank is little although my MAE=0.128 is less than the top rank

👉 This code will take approximately12+ hours on a GPU.Approach I built a pipeline using EfficientNetV2 as the backbone, incorporating metadata-based feature en... See more

Do anyone have the solar panel challenge image dataset downloaded separately ?If you can share this with me via my gdrive..I need this datasets inside my gdrive... See more

IDfCSo1w.jpg（I haven't trained this picture for a while）Judging by the image, it is definitely more than 16+18+81IDfCSo1w S S-unknown 0 16 [(395, 420.0), (393... See more

My solution combines multi-modal data fusion and robust training techniques to tackle the solar panel counting challenge. Key components include: Model Archite... See more

Hello everyone.I know it's a bit late to ask this, but I looked at the code from StarterNotebook_Lacuna and saw that the 'polygon' feature is not used. I didn't... See more

Dear @AJoel , could you please confirm that the evaluation script for this challenge is indeed using MAE . I have just made some experiment on my best submissio... See more

What’s the trick to get public score of less than 1 ? Any tips and suggestions ?

Hi guys. As no one is sharing some hints or notebook, I believe I wont be able to work on this competition. Also finding good regression for segmented images wi... See more

I cant seem to match the coordinates to properly map the panels or boilers in the image, the coordinates are jsut somewhere in the imageHow can i match the poly... See more

Hey guys!!!I have a question, so in the data tab, it was said there that the data contains 3312 images but when i loaded the data i saw 6791 images in the train... See more

could you please share what model you are using

i understand that the polygons are not perfect but are they too bad ?i think that the polgons for theese image might be really bad some of them are:/kaggle/inpu... See more

Hi,Can anyone share a working starter notebook with us. Saw the great different from people having less than 0.5 score and others having above 1.0 score challen... See more

Are the target labels expected to contain errors? I found an image with ID "IDtWBNxeioDGjwvP." According to the CSV data, it should have 8 boilers and 0 panels,... See more

Is it allowed to manually annotate the images provided in the challenge instead of using the existing annotations?