You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The way we write our docs largely influences what user/our code will look like.
Right now we mostly do
from torchvision.transforms import v2
v2.Resize()
There are different options like import torchvision.transforms.v2 as T
We also need to decide how to import the functionals (In the past we've been rather inconsistent and went for things like TF which is potentially confusing).
We also need to decide whether we import datapoints, datapoints as dp, or just import the objects directly from that namespace like from torchvision.datapoints import BoundingBoxes
This is more about user-facing docs / code than about our internal code, but ideally we would align that as well. Our refactored tests use transforms for the module and I find that annoyingly verbose for zero added clarity.
The text was updated successfully, but these errors were encountered:
The way we write our docs largely influences what user/our code will look like.
Right now we mostly do
There are different options like
import torchvision.transforms.v2 as T
We also need to decide how to import the functionals (In the past we've been rather inconsistent and went for things like TF which is potentially confusing).
We also need to decide whether we import
datapoints
,datapoints as dp
, or just import the objects directly from that namespace likefrom torchvision.datapoints import BoundingBoxes
This is more about user-facing docs / code than about our internal code, but ideally we would align that as well. Our refactored tests use
transforms
for the module and I find that annoyingly verbose for zero added clarity.The text was updated successfully, but these errors were encountered: