Skip to content

fix: meta device init condition #1480

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
akoumpa merged 2 commits into from
Mar 7, 2026
Merged

fix: meta device init condition #1480

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6de8a34
fix
adil-a Mar 7, 2026
e8a7ad9
fix
adil-a Mar 7, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions nemo_automodel/_transformers/auto_model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -258,12 +258,13 @@ def _retry(**override):
# Use meta device initialization when:
# - Not using MegatronFSDPManager or DDPManager (they handle their own initialization)
# - AND either multi-GPU (world_size > 1) or single-GPU custom model (not HF)
# - AND not using quantization (we let HF handle BitsAndBytes; don't init meta device)
# - AND not using quantization (we let HF handle BitsAndBytes/FP8; don't init meta device)
# For non-HF models, native quant config is ignored.
is_meta_device = all(
[
not isinstance(model_wrapper, (MegatronFSDPManager, DDPManager)),
get_world_size_safe() > 1 or not is_hf_model,
quantization_config is None and _hf_native_quant_cfg is None,
quantization_config is None and (_hf_native_quant_cfg is None or not is_hf_model),
]
)
init_ctx = ContextManagers([no_init_weights(), init_empty_weights()]) if is_meta_device else nullcontext()
Expand Down
Loading