Skip to content

fix(timm): Catch 'Unknown model' RuntimeError in the Gemma 3n MobileNetV5 vision encoder #40951

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
merged 3 commits into from
Sep 18, 2025
Merged

fix(timm): Catch 'Unknown model' RuntimeError in the Gemma 3n MobileNetV5 vision encoder #40951

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
deab3a9
fix(timm): Add exception handling for unknown Gemma3n model
harshaljanjani Sep 17, 2025
bdfbe33
nit: Let’s cater to this specific issue
harshaljanjani Sep 17, 2025
018c221
nit: Simplify error handling
harshaljanjani Sep 18, 2025
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
28 changes: 25 additions & 3 deletions src/transformers/models/timm_wrapper/modeling_timm_wrapper.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,6 +55,28 @@ class TimmWrapperModelOutput(ModelOutput):
attentions: Optional[tuple[torch.FloatTensor, ...]] = None


def _create_timm_model_with_error_handling(config: "TimmWrapperConfig", **model_kwargs):
"""
Creates a timm model and provides a clear error message if the model is not found,
suggesting a library update.
"""
try:
model = timm.create_model(
config.architecture,
pretrained=False,
**model_kwargs,
)
return model
except RuntimeError as e:
if "Unknown model" in str(e):
# A good general check for unknown models.
raise ImportError(
f"The model architecture '{config.architecture}' is not supported in your version of timm ({timm.__version__}). "
"Please upgrade timm to a more recent version with `pip install -U timm`."
) from e
raise e


@auto_docstring
class TimmWrapperPreTrainedModel(PreTrainedModel):
main_input_name = "pixel_values"
Expand Down Expand Up @@ -138,7 +160,7 @@ def __init__(self, config: TimmWrapperConfig):
super().__init__(config)
# using num_classes=0 to avoid creating classification head
extra_init_kwargs = config.model_args or {}
self.timm_model = timm.create_model(config.architecture, pretrained=False, num_classes=0, **extra_init_kwargs)
self.timm_model = _create_timm_model_with_error_handling(config, num_classes=0, **extra_init_kwargs)
self.post_init()

@auto_docstring
Expand Down Expand Up @@ -254,8 +276,8 @@ def __init__(self, config: TimmWrapperConfig):
)

extra_init_kwargs = config.model_args or {}
self.timm_model = timm.create_model(
config.architecture, pretrained=False, num_classes=config.num_labels, **extra_init_kwargs
self.timm_model = _create_timm_model_with_error_handling(
config, num_classes=config.num_labels, **extra_init_kwargs
)
self.num_labels = config.num_labels
self.post_init()
Expand Down