splits

tomatotest.py

@@ -12,9 +12,9 @@ class CustomConfig(datasets.BuilderConfig):
 
 class RGBSemanticDepthDataset(GeneratorBasedBuilder):
     BUILDER_CONFIGS = [
-        CustomConfig(name="all", version="1.0.0", description="all"),
-        CustomConfig(name="depth", version="1.0.0", description="depth"),
-        CustomConfig(name="seg", version="1.0.0", description="segmentation"),
+        CustomConfig(name="all", version="1.0.0", description="load both segmentation and depth"),
+        CustomConfig(name="depth", version="1.0.0", description="only load depth"),
+        CustomConfig(name="seg", version="1.0.0", description="only load segmentation"),
     ]    # Configs initialization
     BUILDER_CONFIG_CLASS = CustomConfig
     def _info(self):
@@ -31,7 +31,6 @@ class RGBSemanticDepthDataset(GeneratorBasedBuilder):
         # Reference: https://github.com/dwofk/fast-depth/blob/master/dataloaders/dataloader.py#L8-L13
         f = io.BytesIO(bytes_stream)
         h5f = h5py.File(f, "r")
-        print(type_dataset)
         left_rgb = self._read_jpg(h5f['rgb_left'][:])
         if type_dataset == 'depth':
             right_rgb = self._read_jpg(h5f['rgb_right'][:])
@@ -56,21 +55,32 @@ class RGBSemanticDepthDataset(GeneratorBasedBuilder):
             SplitGenerator(
                 name=Split.TRAIN,
                 gen_kwargs={
-                    "archives": [dl_manager.iter_archive(archive) for archive in archives["train"]],
+                    "archives": [dl_manager.iter_archive(archive) for archive in archives],
+                    "split_txt": "train.txt" 
+                },
+            ),
+            SplitGenerator(
+                name=Split.VAL,
+                gen_kwargs={
+                    "archives": [dl_manager.iter_archive(archive) for archive in archives],
+                    "split_txt": "val.txt" 
                 },
             ),
         ]
 
-    def _generate_examples(self, archives):
-        print(self.config.dataset_type)
-        for archive in archives:
-            for path, file in archive:
-                #print(path)
-                left_rgb, right_rgb, left_seg, left_depth, right_depth = self._h5_loader(file.read(), self.config.dataset_type)
-                yield path, {
-                    "left_rgb": left_rgb,
-                    "right_rgb": right_rgb,
-                    "left_seg": left_seg,
-                    "left_depth": left_depth,
-                    "right_depth": right_depth,
-                }
+    def _generate_examples(self, archives, split_txt):
+        with open(split_txt) as split_f:
+            all_splits = split_f.read().split('\n')
+            print(all_splits)
+            for archive in archives:
+                for path, file in archive:
+                    if path not in all_splits:
+                        continue
+                    left_rgb, right_rgb, left_seg, left_depth, right_depth = self._h5_loader(file.read(), self.config.dataset_type)
+                    yield path, {
+                        "left_rgb": left_rgb,
+                        "right_rgb": right_rgb,
+                        "left_seg": left_seg,
+                        "left_depth": left_depth,
+                        "right_depth": right_depth,
+                    }