NASA Harvest Field Boundary Detection Challenge
'%3e%3cg id='b'%3e%3cpath id='a' fill='%23E5BE01' d='M116.1 0 35.692 4.699l76.452 25.35L33.26 13.777l67.286 44.273L28.56 21.915l53.534 60.18-60.18-53.534 36.135 71.985L13.777 33.26l16.272 78.884-25.35-76.452L0 116.1-1-1z'/%3e%3cuse xlink:href='%23a' transform='scale(1 -1)'/%3e%3c/g%3e%3cuse xlink:href='%23b' transform='scale(-1 1)'/%3e%3ccircle r='34.3' fill='%23E5BE01' stroke='%2300A1DE' stroke-width='3.4'/%3e%3c/g%3e%3c/svg%3e)
Helping Rwanda
$5 000 USD
Challenge completed over 2 years ago
Classification
Earth Observation
768 joined
185 active
Start
Nov 17, 22
Close
Feb 26, 23
Reveal
Feb 26, 23
Mix-up with submission format
Data · 29 Dec 2022, 19:27 · 5
Just to prevent anyone the same mistake I made for a couple early submissions, would recommend confirming that row column convention is consistent.
For example:
for i in range(256):
for j in range(256):
f"Tile{tile}_{j}_{i}" = prediction[i][j]
Was correct for me. I'm a little suprised by this since in a numpy array I would have thought i = rows and j = columns, but it seems to be the opposite here.
This would be the proper way to submit it I believe...
Hope this helps somebody!
Thanks.
Can you please, elaborate on the meaning and importance of using the above code. was it that it helps your local cv score to be similar to leaderboard score?
The code above does not necessarily matter depending on your approach, all I'm saying it so make sure your submission is properly assigning rows to row, and columns to columns. If your result is a lot worse than you may have expected after submitting, it's possible you had them mixed up as I did, so I recommend double checking to confirm.
I'm not playing this atm, so perhaps I don't know what I am talking about, but, usually, column first is used for images. So an array would be indexed
Not sure if this clarifies or changes anything?
The way you have it is correct :-) My post was meant to show that I had it the other way around, and if others are also concerned their score may be lower than expected, that was likely one reason.