-
Notifications
You must be signed in to change notification settings - Fork 298
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
File rename: nwpu.py -> vhr10.py #1030
Conversation
I would undo the last commit. It's still called the NWPU VHR-10 dataset. |
If it is called the NWPU VHR-10 dataset then it should be in |
@ashnair1 do you want to comment?
RESISC45 is always referred to as RESISC45 in the literature, but NWPU VHR-10 is always referred to as NWPU VHR-10 in the literature [1], [2]. |
I obviously don't feel very strongly about this (I was the one who chose the original name) but I think @ashnair1 wanted the rename: #847 (comment) |
I agree with @calebrob6. |
My only concern is that Very-High-Resolution isn't a very specific dataset name. It's always referred to as NWPU VHR-10 in the literature, so why not call it that? |
I would think
I haven't worked with this dataset but I see papers referring to it as NWPU RESISC45 Ultimately it comes down to this
Personally I prefer VHR10 throughout. But if we want to use NWPU VHR10 and NWPU RESISC45, I suggest using it in only the |
Tensorflow calls it RESISC45 -- https://www.tensorflow.org/datasets/catalog/resisc45. We can just make the decision to call it VHR-10 because noone else has built out a dataloaders ;) |
To summarize this conversation, I think we all agree on the following:
The only remaining questions to be addressed in this PR are:
I propose we let the authors decide. If we look at the original papers that introduce these datasets, the authors call them NWPU-RESISC45 [1] and NWPU VHR-10 [2], [3]. So my personal preference is to add NWPU to the docs only, and my secondary preference would be to remove NWPU from the section headers but keep it in the dataset docstring. Would the latter be a good compromise? |
|
Should be good to go now. |
* File rename: nwpu.py -> vhr10.py * Update more locations * Name change in several other places * Add NWPU only to docstring Co-authored-by: Caleb Robinson <calebrob6@gmail.com>
This rename was originally done in #847, but I asked @ashnair1 to leave it for a follow-up PR so that the diff would be easier to review. This is the follow-up PR.
I agree with @ashnair1, it doesn't make sense to name the file
nwpu.py
. We already have other datasets from NWPU, VHR-10 isn't the only one. And none of these datasets have anything in common, so there's no need to group them.There are no public API changes since users are supposed to import the dataset from
torchgeo.datasets
, nottorchgeo.datasets.nwpu
.