admin/lerobot
1
0
Fork 0
mirror of https://github.com/huggingface/lerobot.git synced 2026-05-18 10:10:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(language): task_aug style + automatic ${task} rephrasing rotation

Browse Source 
Adds task-prompt diversity (Xiao 2022 / CAST) without touching
``meta/tasks.parquet`` or forcing recipes to opt in. The plan reserved
``task_aug`` as a future style; this lands it now.

- ``language.py``: add ``task_aug`` to ``CORE_STYLES`` and
  ``PERSISTENT_STYLES``. ``column_for_style("task_aug")`` returns
  ``language_persistent`` so PR 2 writers route it correctly.

- ``language_render.py``: ``_resolve_task`` now consults the persistent
  slice for rows of ``style="task_aug", role="user"``. When any exist
  it picks one deterministically by ``sample_idx`` (blake2b-keyed, not
  Python's randomized hash) so an epoch sees every rephrasing of every
  episode while the same sample still resolves identically across
  reruns. Falls back to the canonical ``meta/tasks.parquet`` task when
  no rephrasings are present, so existing datasets and unannotated runs
  keep their behaviour. Explicit ``task=`` overrides still win.

- Tests: rephrasing coverage across samples, determinism on repeat
  ``sample_idx``, fallback when persistent has no ``task_aug`` rows,
  and explicit override priority.

Recipes get this for free: any ``${task}`` placeholder rotates through
the available rephrasings. Recipes that want the literal canonical task
can override the binding.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Pepijn
2026-04-30 16:45:39 +02:00
parent 1ca38d9748
commit c1a0c601e2
4 changed files with 142 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/datasets/test_language.py
+2 -1
View File
@@ -43,7 +43,7 @@ def test_language_arrow_schema_has_expected_fields():
def test_style_registry_routes_columns():
assert {"subtask", "plan", "memory", "motion"} == PERSISTENT_STYLES
assert {"subtask", "plan", "memory", "motion", "task_aug"} == PERSISTENT_STYLES
assert {"interjection", "vqa", "trace"} == EVENT_ONLY_STYLES
assert PERSISTENT_STYLES | EVENT_ONLY_STYLES <= STYLE_REGISTRY
@@ -51,6 +51,7 @@ def test_style_registry_routes_columns():
assert column_for_style("plan") == LANGUAGE_PERSISTENT
assert column_for_style("memory") == LANGUAGE_PERSISTENT
assert column_for_style("motion") == LANGUAGE_PERSISTENT
assert column_for_style("task_aug") == LANGUAGE_PERSISTENT
assert column_for_style("interjection") == LANGUAGE_EVENTS
assert column_for_style("vqa") == LANGUAGE_EVENTS
assert column_for_style("trace") == LANGUAGE_EVENTS