Pepijn e240305e8e fix(annotate): default transformers backend to manual GPU placement ...

Loading Qwen3-VL via transformers + accelerate's device_map='auto'
fails with std::bad_alloc on hosts with abundant RAM. The bug is in
accelerate's post-load dispatch path. Bypassing accelerate by loading
to CPU first and then calling .to('cuda') manually avoids that path.

LEROBOT_TRANSFORMERS_DEVICE_MAP=auto switches back to the old behavior
for cases where it works.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-04-30 18:48:34 +02:00

..

lerobot

fix(annotate): default transformers backend to manual GPU placement

2026-04-30 18:48:34 +02:00