admin/lerobot
1
0
Fork 0
mirror of https://github.com/huggingface/lerobot.git synced 2026-05-26 14:09:47 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(performance): improving implementation for better performances on big datasets

Browse Source
This commit is contained in:
CarolinePascal
2026-05-07 17:50:59 +02:00
parent 6bc94d6e22
commit dd920bf9e3
1 changed files with 4 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/lerobot/datasets/dataset_metadata.py
+4 -3
View File
@@ -205,11 +205,12 @@ class LeRobotDatasetMetadata:
List of sorted episode indices that satisfy the predicate. List of sorted episode indices that satisfy the predicate.
""" """
self.ensure_readable() self.ensure_readable()
ep_table = self.episodes
if candidates is not None: if candidates is not None:
candidate_set = set(candidates) candidate_set = set(candidates)
ep_table = ep_table.filter(lambda ep: ep["episode_index"] in candidate_set) combined = lambda ep: ep["episode_index"] in candidate_set and predicate(ep) # noqa: E731
filtered = ep_table.filter(predicate) else:
combined = predicate
filtered = self.episodes.filter(combined, keep_in_memory=True, load_from_cache_file=False)
return sorted(int(idx) for idx in filtered["episode_index"]) return sorted(int(idx) for idx in filtered["episode_index"])
def _pull_from_repo( def _pull_from_repo(