Skip to content

Instantly share code, notes, and snippets.

@rasha-salim

rasha-salim/creating_dataset.py

Created February 7, 2021 08:45
Show Gist options
  • Save rasha-salim/d1e0c838bfdfbee144f3c5d48fe12dda to your computer and use it in GitHub Desktop.
Save rasha-salim/d1e0c838bfdfbee144f3c5d48fe12dda to your computer and use it in GitHub Desktop.
Download ZIP
Createing Dataset object in activeloop hub
Raw
creating_dataset.py
import hub
from hub.schema import Image
from hub.schema import Tensor
from hub.schema import Mask
from hub.schema import Segmentation
# include your user anme and a name for your dataset
tag = "rasha/landCoverNet_Omdena_Sample"
ds = {}
# Define youe dataset object
ds = Dataset(
tag,
shape=(2512,), # The size of our dataset
mode = "w",
schema = { # Here we'll define the structure of our data
"inputs": Tensor((256, 256, 4), dtype="float32"), # The input images (this is a tif image with 4 bands)
"rgb_inputs": Image(shape=(256, 256, 3), # The RGB version of our input
dtype="uint8"),
"masks": Segmentation(shape=(256, 256, 6), dtype="uint8") # The land cover segmentations
},
)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment