GLC23: added code documentation, cleaned class arguments, updated exa…

…mples consequently.
plantnet · Feb 22, 2024 · a37d163 · a37d163
1 parent 9d49f36
commit a37d163
Show file tree

Hide file tree

Showing 3 changed files with 317 additions and 74 deletions.
diff --git a/examples/kaggle/geolifeclef2023/cnn_on_rgbnir_glc23_patches.py b/examples/kaggle/geolifeclef2023/cnn_on_rgbnir_glc23_patches.py
@@ -12,6 +12,7 @@
 import pytorch_lightning as pl
 from omegaconf import DictConfig
 from pytorch_lightning.callbacks import ModelCheckpoint
+from torchgeo.samplers import Units
 from torchvision import transforms
 
 from malpolon.data.data_module import BaseDataModule
@@ -21,7 +22,6 @@
 from malpolon.logging import Summary
 from malpolon.models import ClassificationSystem
 from malpolon.models.utils import CrashHandler
-from torchgeo.samplers import Units
 
 
 class Sentinel2PatchesDataModule(BaseDataModule):
@@ -94,7 +94,6 @@ def get_dataset(self, split, transform, target_transform=None, **kwargs):
                 providers=[jpp_rgbnir],
                 transform=transform,
                 target_transform=target_transform,
-                id_getitem='patchID',
                 item_columns=['lat', 'lon', 'patchID'],
                 **kwargs
             )

diff --git a/examples/kaggle/geolifeclef2023/example_patch_loading.py b/examples/kaggle/geolifeclef2023/example_patch_loading.py
@@ -28,15 +28,14 @@ def main():
     # create dataset
     dataset = PatchesDataset(occurrences=data_path + 'Presence_only_occurrences/Presences_only_train_sample.csv',
                              providers=[p_hfp_d, p_bioclim, p_hfp_s, p_rgb],
-                             item_columns=['lat', 'lon', 'patchID'],
-                             id_getitem='patchID')
+                             item_columns=['lat', 'lon', 'patchID'])
     dataset_multi = PatchesDatasetMultiLabel(occurrences=data_path + 'Presence_only_occurrences/Presences_only_train_sample.csv',
                                              providers=[p_hfp_d, p_bioclim, p_hfp_s, p_rgb],
                                              item_columns=['lat', 'lon', 'patchID'],
                                              id_getitem='patchID')
 
     # print random tensors from dataset
-    ids = [random.randint(0, len(dataset) - 1) for i in range(1)]
+    ids = [random.randint(0, len(dataset) - 1) for i in range(5)]
     for i in ids:
         tensor, label = dataset[i]
         label_multi = dataset_multi[i][1]