huggingface/pytorch-image-models

Pattern Findings

1509 matches across 14 categories. Click a row to expand file-level details.

Cross-File Repetition243 hits · 1215 pts

Severity	File	Line	Snippet
HIGH	tests/test_models.py	0	run a single forward pass with each model in feature extraction mode
HIGH	tests/test_models.py	0	run a single forward pass with each model in feature extraction mode
HIGH	tests/test_models.py	0	run a single forward pass with each model in feature extraction mode
HIGH	timm/layers/activations.py	0	applies the gaussian error linear units function (w/ dummy inplace arg)
HIGH	timm/layers/activations.py	0	applies the gaussian error linear units function (w/ dummy inplace arg)
HIGH	timm/layers/activations.py	0	applies the gaussian error linear units function (w/ dummy inplace arg)
HIGH	timm/layers/norm_act.py	0	rmsnorm + activation for '2d' nchw tensors note: it's currently (2025-05-10) faster to use an eager 2d kernel that does
HIGH	timm/layers/norm_act.py	0	rmsnorm + activation for '2d' nchw tensors note: it's currently (2025-05-10) faster to use an eager 2d kernel that does
HIGH	timm/layers/norm_act.py	0	rmsnorm + activation for '2d' nchw tensors note: it's currently (2025-05-10) faster to use an eager 2d kernel that does
HIGH	timm/layers/norm_act.py	0	rmsnorm + activation for '2d' nchw tensors note: it's currently (2025-05-10) faster to use an eager 2d kernel that does
HIGH	timm/optim/mars.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/lamb.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/adafactor.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/nvnovograd.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/adabelief.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/adamw.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/laprop.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/nadam.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/rmsprop_tf.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/madgrad.py	0	performs a single optimization step. arguments: closure (callable, optional): a closure that reevaluates the model and r
HIGH	timm/optim/sgdw.py	0	performs a single optimization step. args: closure (callable, optional): a closure that reevaluates the model and return
HIGH	timm/optim/lars.py	0	performs a single optimization step. args: closure (callable, optional): a closure that reevaluates the model and return
HIGH	timm/optim/nadamw.py	0	performs a single optimization step. args: closure (callable, optional): a closure that reevaluates the model and return
HIGH	timm/models/repvit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/tiny_vit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/ghostnet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/davit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/efficientvit_mit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/efficientvit_mit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/rexnet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/edgenext.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/swiftformer.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/nest.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/regnet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/resnetv2.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/efficientvit_msra.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/shvit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/tresnet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/fastvit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/fasternet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/efficientformer.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/levit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/vovnet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/inception_v4.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/maxxvit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/metaformer.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/repghost.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/swin_transformer_v2.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/pvt_v2.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/mambaout.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/gcvit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/starnet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/inception_next.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/focalnet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/pit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/swin_transformer_v2_cr.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/efficientformer_v2.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/nextvit.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/hgnet.py	0	forward features that returns intermediates. args: x: input image tensor indices: take last n blocks if int, all if none
HIGH	timm/models/repvit.py	0	prune layers not required for specified intermediates.
183 more matches not shown…

Unused Imports773 hits · 533 pts

Severity	File	Line
LOW	legacy_train.py	53
LOW	train.py	60
LOW	inference.py	9
LOW	clean_checkpoint.py	15
LOW	tests/test_scheduler.py	18
LOW	tests/test_models.py	30
LOW	timm/__init__.py	1
LOW	timm/__init__.py	2
LOW	timm/__init__.py	2
LOW	timm/__init__.py	2
LOW	timm/__init__.py	2
LOW	timm/__init__.py	8
LOW	timm/__init__.py	8
LOW	timm/__init__.py	8
LOW	timm/__init__.py	8
LOW	timm/__init__.py	8
LOW	timm/__init__.py	8
LOW	timm/__init__.py	8
LOW	timm/__init__.py	8
LOW	timm/__init__.py	8
LOW	timm/layers/squeeze_excite.py	13
LOW	timm/layers/classifier.py	11
LOW	timm/layers/create_act.py	4
LOW	timm/layers/create_act.py	4
LOW	timm/layers/create_act.py	6
LOW	timm/layers/create_act.py	7
LOW	timm/layers/create_norm.py	9
LOW	timm/layers/separable_conv.py	8
LOW	timm/layers/trace_utils.py	2
LOW	timm/layers/__init__.py	1
LOW	timm/layers/__init__.py	1
LOW	timm/layers/__init__.py	1
LOW	timm/layers/__init__.py	1
LOW	timm/layers/__init__.py	1
LOW	timm/layers/__init__.py	1
LOW	timm/layers/__init__.py	1
LOW	timm/layers/__init__.py	1
LOW	timm/layers/__init__.py	11
LOW	timm/layers/__init__.py	12
LOW	timm/layers/__init__.py	12
LOW	timm/layers/__init__.py	12
LOW	timm/layers/__init__.py	12
LOW	timm/layers/__init__.py	18
LOW	timm/layers/__init__.py	18
LOW	timm/layers/__init__.py	18
LOW	timm/layers/__init__.py	18
LOW	timm/layers/__init__.py	19
LOW	timm/layers/__init__.py	19
LOW	timm/layers/__init__.py	19
LOW	timm/layers/__init__.py	20
LOW	timm/layers/__init__.py	20
LOW	timm/layers/__init__.py	21
LOW	timm/layers/__init__.py	21
LOW	timm/layers/__init__.py	22
LOW	timm/layers/__init__.py	22
LOW	timm/layers/__init__.py	23
LOW	timm/layers/__init__.py	23
LOW	timm/layers/__init__.py	23
LOW	timm/layers/__init__.py	23
LOW	timm/layers/__init__.py	24
713 more matches not shown…

Hyper-Verbose Identifiers230 hits · 231 pts

Severity	File	Line	Snippet
LOW	tests/test_utils.py	66	def test_activation_stats_hook_validation():
LOW	tests/test_utils.py	99	def test_freeze_unfreeze_bn_root():
LOW	tests/test_utils.py	116	def test_activation_stats_functions():
LOW	tests/test_utils.py	135	def test_reparameterize_model():
LOW	tests/test_utils.py	164	def test_get_state_dict_custom_unwrap():
LOW	tests/test_utils.py	182	def test_freeze_unfreeze_string_input():
LOW	tests/test_task.py	64	def _param_group_weight_decay(optimizer):
LOW	tests/test_task.py	72	def test_task_checkpoint_omits_empty_task_state_and_keeps_legacy_paths(tmp_path):
LOW	tests/test_task.py	103	def test_feature_distillation_checkpoint_keeps_projection_in_task_state(tmp_path):
LOW	tests/test_task.py	141	def test_feature_distillation_trainable_module_optimizer_grouping():
LOW	tests/test_task.py	167	def test_base_eval_model_preserves_trainable_wrapper_for_use():
LOW	tests/test_task.py	177	def test_compiled_ema_eval_model_reflects_update():
LOW	tests/test_task.py	202	def test_compiled_eval_checkpoint_load_uses_unwrapped_state_dict_target():
LOW	tests/test_layers.py	81	def test_get_act_layer_empty_string():
LOW	tests/test_layers.py	86	def test_create_act_layer_inplace_error():
LOW	tests/test_layers.py	98	def test_create_act_layer_edge_cases():
LOW	tests/test_optim.py	29	def _test_basic_cases_template(weight, bias, input, constructor, scheduler_constructors):
LOW	tests/test_optim.py	289	def _build_params_dict_single(weight, bias, **kwargs):
LOW	tests/test_optim.py	563	def test_param_groups_layer_decay_with_min():
LOW	tests/test_optim.py	586	def test_param_groups_layer_decay_with_matcher():
LOW	tests/test_optim.py	612	def test_param_groups_weight_decay():
LOW	tests/test_scheduler.py	70	def test_plateau_scheduler_step(self):
LOW	tests/test_scheduler.py	122	def test_warmup_prefix_reaches_target_lr(self, scheduler_cls, kwargs):
LOW	tests/test_scheduler.py	347	def test_state_dict_save_load(self, scheduler_cls, kwargs):
LOW	tests/test_scheduler.py	368	def test_plateau_state_dict_save_load(self):
LOW	tests/test_scheduler.py	400	def test_step_update_with_t_in_epochs_false(self, scheduler_cls, kwargs):
LOW	tests/test_scheduler.py	430	def test_multiple_param_groups(self, scheduler_cls, kwargs):
LOW	tests/test_models.py	341	def test_model_default_cfgs_non_std(model_name, batch_size):
LOW	tests/test_models.py	412	def test_model_load_pretrained(model_name, batch_size):
LOW	tests/test_models.py	421	def test_model_features_pretrained(model_name, batch_size):
LOW	tests/test_models.py	431	def test_model_forward_torchscript(model_name, batch_size):
LOW	tests/test_models.py	461	def test_model_forward_features(model_name, batch_size):
LOW	tests/test_models.py	492	def test_model_forward_intermediates_features(model_name, batch_size):
LOW	tests/test_models.py	523	def test_model_forward_intermediates(model_name, batch_size):
LOW	tests/test_models.py	712	def test_model_forward_fx_torchscript(model_name, batch_size):
LOW	tests/test_models.py	737	def test_model_forward_torchscript_with_features_fx(model_name, batch_size):
LOW	tests/test_checkpoint_loading.py	23	def test_weights_only_allows_argparse_namespace(tmp_path):
LOW	tests/test_checkpoint_loading.py	36	def test_weights_only_blocks_non_allowlisted_globals(tmp_path):
LOW	tests/test_checkpoint_loading.py	52	def test_resume_checkpoint_default_weights_only_namespace(tmp_path):
LOW	tests/test_checkpoint_loading.py	82	def test_resume_checkpoint_blocks_non_allowlisted_globals(tmp_path):
LOW	tests/test_checkpoint_loading.py	95	def test_resume_checkpoint_weights_only_false_allows_custom_globals(tmp_path):
LOW	tests/test_layers_drop.py	125	def test_deprecated_args_accepted(self):
LOW	tests/test_layers_drop.py	156	def test_couple_channels_parameter(self):
LOW	tests/test_layers_drop.py	186	def test_no_drop_when_prob_zero(self):
LOW	tests/test_layers_drop.py	192	def test_no_drop_when_not_training(self):
LOW	tests/test_layers_drop.py	214	def test_drop_path_no_scaling(self):
LOW	tests/test_layers_pool.py	78	def test_attention_pool_latent_basic(self):
LOW	tests/test_layers_pool.py	85	def test_attention_pool_latent_multi_latent(self):
LOW	tests/test_layers_pool.py	231	def test_attention_pool_prr_basic(self):
LOW	tests/test_layers_pool.py	238	def test_attention_pool_prr_avg_pool(self):
LOW	tests/test_layers_pool.py	245	def test_attention_pool_prr_parameter_free(self):
LOW	tests/test_layers_pool.py	251	def test_attention_pool_prr_with_norms(self):
LOW	tests/test_layers_pool.py	274	def test_attention_pool_latent_out_features(self, out_features, embed_dim, expected_out):
LOW	tests/test_layers_pool.py	322	def test_lse_high_r_approximates_max(self):
LOW	tests/test_layers_pool.py	330	def test_lse_low_r_approximates_avg(self):
LOW	tests/test_layers_pool.py	338	def test_lse_learnable_r_gradient(self):
LOW	tests/test_layers_pool.py	460	def test_different_spatial_sizes(self, pool_cls, kwargs, input_shape):
LOW	timm/layers/pos_embed_sincos.py	200	def init_non_persistent_buffers(self) -> None:
LOW	timm/layers/pos_embed_sincos.py	496	def init_non_persistent_buffers(self) -> None:
LOW	timm/layers/pos_embed_sincos.py	630	def init_non_persistent_buffers(self) -> None:
170 more matches not shown…

Deep Nesting149 hits · 149 pts

Severity	File	Line
LOW	legacy_train.py	428
LOW	legacy_train.py	1076
LOW	legacy_train.py	1288
LOW	benchmark.py	163
LOW	benchmark.py	564
LOW	benchmark.py	620
LOW	benchmark.py	305
LOW	validate.py	175
LOW	validate.py	462
LOW	validate.py	491
LOW	train.py	460
LOW	train.py	1176
LOW	train.py	1392
LOW	bulk_runner.py	96
LOW	bulk_runner.py	141
LOW	inference.py	150
LOW	inference.py	353
LOW	results/generate_csv_results.py	20
LOW	timm/layers/classifier.py	282
LOW	timm/layers/pos_embed_sincos.py	1235
LOW	timm/layers/gather_excite.py	29
LOW	timm/layers/non_local_attn.py	71
LOW	timm/layers/adaptive_avgmax_pool.py	42
LOW	timm/layers/adaptive_avgmax_pool.py	127
LOW	timm/layers/inplace_abn.py	43
LOW	timm/layers/pool1d.py	4
LOW	timm/layers/create_attn.py	22
LOW	timm/layers/weight_init.py	110
LOW	timm/layers/norm_act.py	278
LOW	timm/layers/norm_act.py	322
LOW	timm/scheduler/scheduler_factory.py	63
LOW	timm/optim/_param_groups.py	19
LOW	timm/optim/sgdw.py	195
LOW	timm/optim/sgdw.py	241
LOW	timm/optim/_optim_factory.py	177
LOW	timm/optim/lamb.py	153
LOW	timm/optim/adafactor.py	128
LOW	timm/optim/adafactor_bv.py	210
LOW	timm/optim/adafactor_bv.py	123
LOW	timm/optim/lars.py	76
LOW	timm/optim/nvnovograd.py	67
LOW	timm/optim/adamp.py	90
LOW	timm/optim/adabelief.py	107
LOW	timm/optim/adamw.py	103
LOW	timm/optim/kron.py	400
LOW	timm/optim/kron.py	218
LOW	timm/optim/radam.py	38
LOW	timm/optim/laprop.py	80
LOW	timm/optim/muon.py	531
LOW	timm/optim/muon.py	783
LOW	timm/optim/rmsprop_tf.py	100
LOW	timm/optim/sgdp.py	50
LOW	timm/optim/madgrad.py	91
LOW	timm/utils/model.py	111
LOW	timm/utils/model.py	233
LOW	timm/utils/model.py	237
LOW	timm/models/davit.py	470
LOW	timm/models/inception_resnet_v2.py	299
LOW	timm/models/inception_resnet_v2.py	303
LOW	timm/models/hrnet.py	439
89 more matches not shown…

Docstring Block Structure11 hits · 55 pts

Severity	File	Line	Snippet
HIGH	timm/optim/_optim_factory.py	161	Get the OptimInfo for an optimizer. Args: name: Name of the optimizer Returns:
HIGH	timm/optim/_optim_factory.py	182	Get the optimizer class with any default arguments applied. This allows direct instantiation of optimizers with
HIGH	timm/optim/_optim_factory.py	245	Create an optimizer instance. Args: model_or_params: Model or parameters to optimize op
HIGH	timm/optim/_optim_factory.py	1107	List available optimizer names, optionally filtered. List all registered optimizers, with optional filtering using
HIGH	timm/optim/_optim_factory.py	1148	Get the OptimInfo for an optimizer. Args: name: Name of the optimizer Returns: OptimInfo confi
HIGH	timm/optim/_optim_factory.py	1166	Get optimizer class by name with option to bind default arguments. Retrieves the optimizer class or a partial funct
HIGH	timm/optim/_optim_factory.py	1215	Create an optimizer instance via timm registry. Creates and configures an optimizer with appropriate parameter grou
HIGH	timm/optim/muon.py	267	Check if a parameter is suitable for Muon optimization. Args: param: Parameter tensor min_dim_size:
HIGH	timm/models/_efficientnet_builder.py	82	Decode block definition string Gets a list of block arg (dicts) through a string notation of arguments. E.g. i
HIGH	timm/data/auto_augment.py	587	Create a AutoAugment transform Args: config_str: String defining configuration of auto augmentation. Consi
HIGH	timm/data/auto_augment.py	767	Create a RandAugment transform Args: config_str (str): String defining configuration of random augmentatio

Decorative Section Separators15 hits · 45 pts

Severity	File	Line	Snippet
MEDIUM	timm/optim/lion.py	22	# ==============================================================================
MEDIUM	timm/models/beit.py	29	# --------------------------------------------------------
MEDIUM	timm/models/twins.py	8	# --------------------------------------------------------
MEDIUM	timm/models/twins.py	13	# --------------------------------------------------------
MEDIUM	timm/models/swin_transformer_v2.py	9	# --------------------------------------------------------
MEDIUM	timm/models/swin_transformer_v2.py	14	# --------------------------------------------------------
MEDIUM	timm/models/swin_transformer.py	12	# --------------------------------------------------------
MEDIUM	timm/models/swin_transformer.py	17	# --------------------------------------------------------
MEDIUM	timm/models/focalnet.py	14	# --------------------------------------------------------
MEDIUM	timm/models/focalnet.py	19	# --------------------------------------------------------
MEDIUM	timm/models/hiera.py	11	# --------------------------------------------------------
MEDIUM	timm/models/hiera.py	24	# --------------------------------------------------------
MEDIUM	timm/models/swin_transformer_v2_cr.py	24	# --------------------------------------------------------
MEDIUM	timm/models/swin_transformer_v2_cr.py	29	# --------------------------------------------------------
MEDIUM	timm/data/tf_preprocessing.py	23	# ==============================================================================

Hallucination Indicators4 hits · 40 pts

Severity	File	Line	Snippet
CRITICAL	timm/models/hieradet_sam2.py	426	self.head.fc.weight.data.mul_(head_init_scale)
CRITICAL	timm/models/hieradet_sam2.py	427	self.head.fc.bias.data.mul_(head_init_scale)
CRITICAL	timm/models/hiera.py	631	self.head.fc.weight.data.mul_(head_init_scale)
CRITICAL	timm/models/hiera.py	632	self.head.fc.bias.data.mul_(head_init_scale)

Over-Commented Block28 hits · 28 pts

Severity	File	Line	Snippet
LOW	tests/test_optim.py	341	# _test_basic_cases(
LOW	timm/layers/pos_embed_rel.py	41
LOW	timm/layers/patch_embed.py	601	#
LOW	timm/layers/patch_embed.py	621	# self.patch_size = to_2tuple(patch_size)
LOW	timm/layers/patch_embed.py	641	# else:
LOW	timm/layers/cond_conv2d.py	121
LOW	timm/optim/lion.py	1	""" Lion Optimizer
LOW	timm/optim/lamb.py	21	# Copyright (c) 2021, Habana Labs Ltd. All rights reserved.
LOW	timm/optim/lamb.py	41	# Permission is hereby granted, free of charge, to any person obtaining a copy
LOW	timm/optim/adan.py	1	""" Adan Optimizer
LOW	timm/utils/distributed.py	121	dist_backend = dist_backends.get(device_type, 'gloo')
LOW	timm/models/volo.py	1	""" Vision OutLOoker (VOLO) implementation
LOW	timm/models/efficientvit_mit.py	1161	hf_hub_id='timm/',
LOW	timm/models/efficientvit_mit.py	1261	# def efficientvit_l0_sam(pretrained=False, **kwargs):
LOW	timm/models/beit.py	21	primaryClass={cs.CV}
LOW	timm/models/efficientvit_msra.py	661
LOW	timm/models/efficientvit_msra.py	681	# for k, v in tmp_dict.items():
LOW	timm/models/shvit.py	461
LOW	timm/models/shvit.py	481	# downsample_mapping[f'^stages\\.{i}\\.blocks\\.{j}\\.'] = f'stages.{i}.blocks.{j - 3}.'
LOW	timm/models/fasternet.py	401	# }
LOW	timm/models/twins.py	101	x = x[:, :H, :W, :].contiguous()
LOW	timm/models/twins.py	121	# x = x.reshape(B, _h, self.ws, _w, self.ws, C).transpose(2, 3) # B, _h, _w, ws, ws, C
LOW	timm/models/visformer.py	521	attn_stage='011', spatial_conv='100', norm_layer=nn.BatchNorm2d, conv_init=True,
LOW	timm/models/visformer.py	541	# model.default_cfg = _cfg()
LOW	timm/models/visformer.py	561	#
LOW	timm/models/visformer.py	581	# @register_model
LOW	timm/models/hiera.py	1	""" An PyTorch implementation of Hiera
LOW	timm/data/tf_preprocessing.py	1	""" Tensorflow Preprocessing Adapter

Excessive Try-Catch Wrapping23 hits · 28 pts

Severity	File	Line	Snippet
MEDIUM	avg_checkpoints.py	85	print("Error: Output filename ({}) already exists.".format(output))
MEDIUM	avg_checkpoints.py	113	print('Error: No checkpoints found to average.')
MEDIUM	avg_checkpoints.py	121	print(f"Error: Checkpoint ({c}) doesn't exist")
LOW	bulk_runner.py	203	except Exception as e:
MEDIUM	clean_checkpoint.py	42	print("Error: Output filename ({}) already exists.".format(args.output))
MEDIUM	clean_checkpoint.py	110	print("Error: Checkpoint ({}) doesn't exist".format(checkpoint))
LOW	tests/test_models.py	199	except Exception:
LOW	tests/test_models.py	560	except Exception:
LOW	timm/layers/blur_pool.py	118	except Exception:
LOW	timm/optim/lion.py	159	except Exception:
LOW	timm/optim/adan.py	127	except Exception:
MEDIUM	timm/optim/adan.py	34	def __init__(self, chunk_size):
LOW	timm/optim/adamw.py	216	except Exception:
LOW	timm/optim/nadamw.py	196	except Exception:
LOW	timm/utils/jit.py	33	except Exception:
LOW	timm/utils/jit.py	42	except Exception:
LOW	timm/utils/checkpoint_saver.py	122	except Exception as e:
LOW	timm/utils/checkpoint_saver.py	178	except Exception as e:
LOW	timm/models/eva.py	1228	except Exception as e:
LOW	timm/models/_hub.py	572	except Exception as _:
LOW	timm/models/_helpers.py	36	except Exception:
LOW	timm/data/naflex_dataset.py	549	except Exception as e:
LOW	timm/data/readers/reader_wds.py	57	except Exception as e:

Redundant / Tautological Comments16 hits · 24 pts

Severity	File	Line	Snippet
LOW	legacy_train.py	689	# Check if we should use the NaFlex scheduled loader
LOW	legacy_train.py	753	# Check if we have model's patch size for NaFlex mode
LOW	train.py	671	# Check if we should use the NaFlex scheduled loader
LOW	train.py	735	# Check if we have model's patch size for NaFlex mode
LOW	timm/layers/pos_embed_sincos.py	682	# Check if we have pre-computed bands
LOW	timm/layers/blur_pool.py	143	# Check if aa_layer is a partial and already has device/dtype set
LOW	timm/optim/muon.py	576	# Increment step
LOW	timm/models/vision_transformer.py	3072	# Check if we should use NaFlexVit instead
LOW	timm/models/fastvit.py	96	# Check if SE-ReLU is requested
LOW	timm/models/naflexvit.py	281	# Check if we need EVA block features
LOW	timm/models/naflexvit.py	1937	# Check if it's a perfect square for a standard grid
LOW	timm/models/eva.py	1297	# Check if we should use NaFlexVit implementation
LOW	timm/models/_hub.py	419	# Check if README file already exist in repo
LOW	timm/models/_helpers.py	111	# Check if safetensors or not and load weights accordingly
LOW	timm/data/naflex_dataset.py	119	# Check if patches are flattened or unflattened
LOW	convert/convert_gemma4_vit.py	97	# Check if this is a vision key (vision_tower or vision_model)

Self-Referential Comments10 hits · 21 pts

Severity	File	Line	Snippet
MEDIUM	timm/models/_factory.py	85	>>> # Create a MobileNetV3-Large model with no pretrained weights.
MEDIUM	timm/models/_factory.py	88	>>> # Create a MobileNetV3-Large model with pretrained weights.
MEDIUM	timm/models/_factory.py	93	>>> # Create a MobileNetV3-Large model with pretrained weights and a new head with 10 classes.
MEDIUM	timm/models/_factory.py	98	>>> # Create a Dinov2 small model with pretrained weights and save weights in a custom directory.
MEDIUM	timm/models/naflexvit.py	1024	# Create the float mask and apply masking using additive mask convention
MEDIUM	timm/models/hiera.py	154	# Create a view of the tensor with the patch stride as separate dims
MEDIUM	timm/models/swin_transformer_v2_cr.py	98	r"""This class implements window-based Multi-Head-Attention with log-spaced continuous position bias.
MEDIUM	timm/models/swin_transformer_v2_cr.py	246	r"""This class implements the Swin transformer block.
MEDIUM	timm/models/swin_transformer_v2_cr.py	594	r"""This class implements a stage of the Swin transformer including multiple layers.
MEDIUM	timm/data/naflex_loader.py	422	# Create the collator

AI Slop Vocabulary5 hits · 10 pts

Severity	File	Line	Snippet
MEDIUM	timm/layers/norm_act.py	156	# but ONLY when used in conjunction with the timm conversion function below.
MEDIUM	timm/models/gemma4_vit.py	1364	# so the test harness' size-clamp doesn't land at 128 and blow the
MEDIUM	timm/models/swin_transformer_v2_cr.py	681	""" Updates the resolution to utilize and the window size and so the pair-wise relative positions.
MEDIUM	timm/data/imagenet_info.py	41	# which will be more robust than a guess based on num_classes
LOW	timm/data/naflex_dataset.py	61	# If no divisor specified, just use integer division

Slop Phrases1 hit · 2 pts

Severity	File	Line	Snippet
LOW	timm/models/xception.py	286	# The resize parameter of the validation transform should be 333, and make sure to center crop at 299x299

Verbosity Indicators1 hit · 2 pts

Severity	File	Line	Snippet
LOW	timm/data/naflex_transforms.py	261	# Step 1: Get the maximum allowed dimensions from sequence length constraint

Score History

Severity Breakdown

Pattern Findings