Correcting recent bugs code after debugging on devfair.

Summary:
1. Typo in the dataset path in the config.
2. Typo in num_frames.
3. Pick sequence was cached before it was modified for single-sequence.

Reviewed By: bottler

Differential Revision: D36417329

fbshipit-source-id: 6dcd75583de510412e1ae58f63db04bb4447403e

projects/implicitron_trainer/configs/repro_base.yaml

@@ -21,7 +21,7 @@ dataloader_args:
   - 9
   - 10
 dataset_args:
-  dataset_root: ${oc.env:CO3D_DATASET_ROOT}"
+  dataset_root: ${oc.env:CO3D_DATASET_ROOT}
   load_point_clouds: false
   mask_depths: false
   mask_images: false

pytorch3d/implicitron/dataset/dataset_zoo.py

@@ -146,7 +146,6 @@ def dataset_zoo(
             "load_point_clouds": load_point_clouds,
             "mask_images": mask_images,
             "mask_depths": mask_depths,
-            "pick_sequence": restrict_sequence_name,
             "path_manager": path_manager,
             "frame_annotations_file": frame_file,
             "sequence_annotations_file": sequence_file,
@@ -174,7 +173,10 @@ def dataset_zoo(
         if not os.path.isfile(batch_indices_path):
             # The batch indices file does not exist.
             # Most probably the user has not specified the root folder.
-            raise ValueError("Please specify a correct dataset_root folder.")
+            raise ValueError(
+                f"Looking for batch indices in {batch_indices_path}. "
+                + "Please specify a correct dataset_root folder."
+            )
 
         with open(batch_indices_path, "r") as f:
             eval_batch_index = json.load(f)
@@ -208,6 +210,7 @@ def dataset_zoo(
             train_dataset = ImplicitronDataset(
                 n_frames_per_sequence=n_frames_per_sequence,
                 subsets=set_names_mapping["train"],
+                pick_sequence=restrict_sequence_name,
                 **common_kwargs,
             )
         if test_on_train:
@@ -215,13 +218,15 @@ def dataset_zoo(
             val_dataset = test_dataset = train_dataset
         else:
             val_dataset = ImplicitronDataset(
-                n_frames_per_sequence=1,
+                n_frames_per_sequence=-1,
                 subsets=set_names_mapping["val"],
+                pick_sequence=restrict_sequence_name,
                 **common_kwargs,
             )
             test_dataset = ImplicitronDataset(
-                n_frames_per_sequence=1,
+                n_frames_per_sequence=-1,
                 subsets=set_names_mapping["test"],
+                pick_sequence=restrict_sequence_name,
                 **common_kwargs,
             )
             if len(restrict_sequence_name) > 0: