What causes "Importing from {__name__} is deprecated, please import via timm.models" in huggingface/pytorch-image-models?

Use from timm.models import FeatureListNet, FeatureDictNet

huggingface/pytorch-image-models · warning · FutureWarning

Importing from {name} is deprecated, please import via t

Error message

Importing from {__name__} is deprecated, please import via timm.models

What it means

Importing timm.models.features is deprecated: the module moved to timm.models._features and this shim re-exports it with a FutureWarning. Import from timm.models (or the explicit timm.models._features) instead.

Source

Thrown at timm/models/features.py:4

from ._features import *

import warnings
warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.models", FutureWarning)

View on GitHub (pinned to 9a5261e31b)

Solutions

Use from timm.models import FeatureListNet, FeatureDictNet
Or from timm.models._features import ... for exact module parity
Run a repo-wide grep for 'timm.models.features' and replace

Example fix

# before
from timm.models.features import FeatureListNet
# after
from timm.models import FeatureListNet

Defensive patterns

Strategy: validation

Validate before calling

# migrate instead of guarding
from timm.models import FeatureListNet  # correct modern import

Prevention

Use from timm.models import ... for feature nets
Grep for 'timm.models.features' in CI
Prefer create_model(..., features_only=True) for feature extraction

When it happens

Trigger: from timm.models.features import FeatureListNet, FeatureDictNet, feature_take_indices; import timm.models.features in code written for older timm.

Common situations: Feature-extraction pipelines (intermediate layer tapping) written against pre-0.9 timm; repos using timm for perceptual losses or FPN heads.

Related errors

AI-assisted analysis of huggingface/pytorch-image-models@9a5261e31b (2026-08-27). Data as JSON: /api/errors/792a7ff7b9ddee53. Report an issue: GitHub.