Implented visualize_image_transforms script

lerobot/scripts/show_image_transforms.py

@@ -1,27 +1,30 @@
 from pathlib import Path
 
+import hydra
+
 from torchvision.transforms import ToPILImage
 
 from lerobot.common.datasets.lerobot_dataset import LeRobotDataset
 from lerobot.common.datasets.transforms import make_transforms
-from lerobot.common.utils.utils import init_hydra_config
 
-DEFAULT_CONFIG_PATH = "lerobot/configs/default.yaml"
 to_pil = ToPILImage()
 
 
-def main(repo_id):
+def main(cfg, output_dir=Path("outputs/image_transforms")):
     """
-    Apply a series of image transformations to a frame from a dataset and save the transformed images.
+    Function to apply image transforms from a configuration and save the transformed images.
 
     Args:
-        repo_id (str): The ID of the repository.
+        cfg (object): Configuration object containing the image transform settings and dataset_repo_id.
+        output_dir (str or Path, optional): Output directory to save the transformed images. Defaults to "outputs/image_transforms".
+
+    Returns:
+        None
     """
     
-    transforms = ["colorjitter", "sharpness", "blur"]
+    dataset = LeRobotDataset(cfg.dataset_repo_id, transform=None)
     
-    dataset = LeRobotDataset(repo_id, transform=None)
+    output_dir = Path(output_dir) / Path(cfg.dataset_repo_id.split("/")[-1])
-    output_dir = Path("outputs/image_transforms") / Path(repo_id)
     output_dir.mkdir(parents=True, exist_ok=True)
 
     # Get first frame of 1st episode
@@ -33,25 +36,25 @@ def main(repo_id):
 
     # Apply each single transformation
     for transform_name in transforms:
-        overrides = [
+        cfg.image_transform.enable=True
-                "image_transform.enable=True",
+        cfg.image_transform.max_num_transforms=1
-                "image_transform.max_num_transforms=1",
+
-        ]
         for t in transforms:
             if t == transform_name:
-                overrides.append(f"image_transform.{t}.weight=1")
+                cfg.image_transform[t].weight=1
-                overrides.append(f"image_transform.{t}_p=1")
             else:
-                overrides.append(f"image_transform.{t}.weight=0")
+                cfg.image_transform[t].weight=0
-        cfg = init_hydra_config(
+
-            DEFAULT_CONFIG_PATH,
-            overrides=overrides,
-        )
         transform = make_transforms(cfg.image_transform)
         img = transform(frame)
         to_pil(img).save(output_dir / f"{transform_name}.png", quality=100)
 
 
+@hydra.main(version_base="1.2", config_name="default", config_path="../configs")
+def visualize_transforms_cli(cfg: dict):
+    main(
+        cfg,
+    )
+
 if __name__ == "__main__":
-    repo_id = "cadene/reachy2_teleop_remi"
+    visualize_transforms_cli()
-    main(repo_id)