fix unit tests, stats was missing, visualize_dataset was broken

lerobot/common/datasets/factory.py

@@ -53,7 +53,11 @@ def make_dataset(
             stats["action"]["max"] = torch.tensor([511.0, 511.0], dtype=torch.float32)
         elif stats_path is None:
             # load stats if the file exists already or compute stats and save it
-            precomputed_stats_path = Path("data") / cfg.dataset_id / "stats.pth"
+            if DATA_DIR is None:
+                # TODO(rcadene): clean stats
+                precomputed_stats_path = Path("data") / cfg.dataset_id / "stats.pth"
+            else:
+                precomputed_stats_path = DATA_DIR / cfg.dataset_id / "stats.pth"
             if precomputed_stats_path.exists():
                 stats = torch.load(precomputed_stats_path)
             else:

lerobot/scripts/download_and_upload_dataset.py

@@ -5,6 +5,7 @@ useless dependencies when using datasets.
 
 import io
 import pickle
+import shutil
 from pathlib import Path
 
 import einops
@@ -44,7 +45,7 @@ def download_and_extract_zip(url: str, destination_folder: Path) -> bool:
         return False
 
 
-def download_and_upload_pusht(root, dataset_id="pusht", fps=10):
+def download_and_upload_pusht(root, root_tests, dataset_id="pusht", fps=10):
     try:
         import pymunk
         from gym_pusht.envs.pusht import PushTEnv, pymunk_to_shapely
@@ -197,12 +198,12 @@ def download_and_upload_pusht(root, dataset_id="pusht", fps=10):
     dataset = dataset.with_format("torch")
 
     num_items_first_ep = ep_dicts[0]["frame_id"].shape[0]
-    dataset.select(range(num_items_first_ep)).save_to_disk(f"tests/data/{dataset_id}/train")
+    dataset.select(range(num_items_first_ep)).save_to_disk(f"{root_tests}/{dataset_id}/train")
     dataset.push_to_hub(f"lerobot/{dataset_id}", token=True)
     dataset.push_to_hub(f"lerobot/{dataset_id}", token=True, revision="v1.0")
 
 
-def download_and_upload_xarm(root, dataset_id, fps=15):
+def download_and_upload_xarm(root, root_tests, dataset_id, fps=15):
     root = Path(root)
     raw_dir = root / f"{dataset_id}_raw"
     if not raw_dir.exists():
@@ -308,12 +309,12 @@ def download_and_upload_xarm(root, dataset_id, fps=15):
     dataset = dataset.with_format("torch")
 
     num_items_first_ep = ep_dicts[0]["frame_id"].shape[0]
-    dataset.select(range(num_items_first_ep)).save_to_disk(f"tests/data/{dataset_id}/train")
+    dataset.select(range(num_items_first_ep)).save_to_disk(f"{root_tests}/{dataset_id}/train")
     dataset.push_to_hub(f"lerobot/{dataset_id}", token=True)
     dataset.push_to_hub(f"lerobot/{dataset_id}", token=True, revision="v1.0")
 
 
-def download_and_upload_aloha(root, dataset_id, fps=50):
+def download_and_upload_aloha(root, root_tests, dataset_id, fps=50):
     folder_urls = {
         "aloha_sim_insertion_human": "https://drive.google.com/drive/folders/1RgyD0JgTX30H4IM5XZn8I3zSV_mr8pyF",
         "aloha_sim_insertion_scripted": "https://drive.google.com/drive/folders/1TsojQQSXtHEoGnqgJ3gmpPQR2DPLtS2N",
@@ -453,16 +454,30 @@ def download_and_upload_aloha(root, dataset_id, fps=50):
     dataset = dataset.with_format("torch")
 
     num_items_first_ep = ep_dicts[0]["frame_id"].shape[0]
-    dataset.select(range(num_items_first_ep)).save_to_disk(f"tests/data/{dataset_id}/train")
+    dataset.select(range(num_items_first_ep)).save_to_disk(f"{root_tests}/{dataset_id}/train")
     dataset.push_to_hub(f"lerobot/{dataset_id}", token=True)
     dataset.push_to_hub(f"lerobot/{dataset_id}", token=True, revision="v1.0")
 
 
 if __name__ == "__main__":
     root = "data"
-    download_and_upload_pusht(root, dataset_id="pusht")
+    root_tests = "{root_tests}"
-    download_and_upload_xarm(root, dataset_id="xarm_lift_medium")
+
-    download_and_upload_aloha(root, dataset_id="aloha_sim_insertion_human")
+    download_and_upload_pusht(root, root_tests, dataset_id="pusht")
-    download_and_upload_aloha(root, dataset_id="aloha_sim_insertion_scripted")
+    download_and_upload_xarm(root, root_tests, dataset_id="xarm_lift_medium")
-    download_and_upload_aloha(root, dataset_id="aloha_sim_transfer_cube_human")
+    download_and_upload_aloha(root, root_tests, dataset_id="aloha_sim_insertion_human")
-    download_and_upload_aloha(root, dataset_id="aloha_sim_transfer_cube_scripted")
+    download_and_upload_aloha(root, root_tests, dataset_id="aloha_sim_insertion_scripted")
+    download_and_upload_aloha(root, root_tests, dataset_id="aloha_sim_transfer_cube_human")
+    download_and_upload_aloha(root, root_tests, dataset_id="aloha_sim_transfer_cube_scripted")
+
+    dataset_ids = [
+        "pusht",
+        "xarm_lift_medium",
+        "aloha_sim_insertion_human",
+        "aloha_sim_insertion_scripted",
+        "aloha_sim_transfer_cube_human",
+        "aloha_sim_transfer_cube_scripted",
+    ]
+    for dataset_id in dataset_ids:
+        # assume stats have been precomputed
+        shutil.copy(f"{root}/{dataset_id}/stats.pth", f"{root_tests}/{dataset_id}/stats.pth")

lerobot/scripts/visualize_dataset.py

@@ -62,12 +62,12 @@ def render_dataset(dataset, out_dir, max_num_episodes):
     )
     dl_iter = iter(dataloader)
 
-    num_episodes = len(dataset.data_ids_per_episode)
+    for ep_id in range(min(max_num_episodes, dataset.num_episodes)):
-    for ep_id in range(min(max_num_episodes, num_episodes)):
         logging.info(f"Rendering episode {ep_id}")
 
         frames = {}
-        for _ in dataset.data_ids_per_episode[ep_id]:
+        end_of_episode = False
+        while not end_of_episode:
             item = next(dl_iter)
 
             for im_key in dataset.image_keys:
@@ -77,6 +77,8 @@ def render_dataset(dataset, out_dir, max_num_episodes):
                 # add current frame to list of frames to render
                 frames[im_key].append(item[im_key])
 
+            end_of_episode = item["index"].item() == item["episode_data_id_to"].item()
+
         out_dir.mkdir(parents=True, exist_ok=True)
         for im_key in dataset.image_keys:
             if len(dataset.image_keys) > 1: