AttributeError: 'SSDHead' object has no attribute 'loss_cls'

[dpdevi9]

error

Traceback (most recent call last): File "/content/mmdetection/tools/train.py", line 121, in main() File "/content/mmdetection/tools/train.py", line 117, in main runner.train() File "/usr/local/lib/python3.10/dist-packages/mmengine/runner/runner.py", line 1777, in train model = self.train_loop.run() # type: ignore File "/usr/local/lib/python3.10/dist-packages/mmengine/runner/loops.py", line 102, in run self.runner.val_loop.run() File "/usr/local/lib/python3.10/dist-packages/mmengine/runner/loops.py", line 363, in run self.run_iter(idx, data_batch) File "/usr/local/lib/python3.10/dist-packages/torch/utils/_contextlib.py", line 115, in decorate_context return func(args, kwargs) File "/usr/local/lib/python3.10/dist-packages/mmengine/runner/loops.py", line 383, in run_iter outputs = self.runner.model.val_step(data_batch) File "/usr/local/lib/python3.10/dist-packages/mmengine/model/base_model/base_model.py", line 133, in val_step return self._run_forward(data, mode='predict') # type: ignore File "/usr/local/lib/python3.10/dist-packages/mmengine/model/base_model/base_model.py", line 346, in _run_forward results = self(data, mode=mode) File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(args, **kwargs) File "/content/mmdetection/mmdet/models/detectors/base.py", line 94, in forward return self.predict(inputs, data_samples) File "/content/mmdetection/mmdet/models/detectors/single_stage.py", line 110, in predict results_list = self.bbox_head.predict( File "/content/mmdetection/mmdet/models/dense_heads/base_dense_head.py", line 197, in predict predictions = self.predict_by_feat( File "/content/mmdetection/mmdet/models/dense_heads/base_dense_head.py", line 279, in predict_by_feat results = self._predict_by_feat_single( File "/content/mmdetection/mmdet/models/dense_heads/base_dense_head.py", line 376, in _predict_by_feat_single if getattr(self.loss_cls, 'custom_cls_channels', False): File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1614, in getattr raise AttributeError("'{}' object has no attribute '{}'".format( AttributeError: 'SSDHead' object has no attribute 'loss_cls'

Reproduction python tools/train.py configs/ssd/ssd_my.py

ssd_my.py

#########################################################################################################
# MODEL

num_classes = 1
input_size = img_size = 512

model = dict(
    type='SingleStageDetector',
    data_preprocessor=dict(
        type='DetDataPreprocessor',
        mean=[123.675, 116.28, 103.53],
        std=[1, 1, 1],
        bgr_to_rgb=True,
        pad_size_divisor=1),
    backbone=dict(
        type='SSDVGG',
        depth=16,
        with_last_pool=False,
        ceil_mode=True,
        out_indices=(3, 4),
        out_feature_indices=(22, 34),
        init_cfg=dict(
            type='Pretrained', checkpoint='open-mmlab://vgg16_caffe')),
    neck=dict(
        type='SSDNeck',
        in_channels=(512, 1024),
        out_channels=(512, 1024, 512, 256, 256, 256),
        level_strides=(2, 2, 1, 1),
        level_paddings=(1, 1, 0, 0),
        l2_norm_scale=20),
    bbox_head=dict(
        type='SSDHead',
        in_channels=(512, 1024, 512, 256, 256, 256),
        num_classes=1,
        anchor_generator=dict(
            type='SSDAnchorGenerator',
            scale_major=False,
            input_size=input_size,
            basesize_ratio_range=(0.15, 0.9),
            strides=[8, 16, 32, 64, 100, 300],
            ratios=[[2], [2, 3], [2, 3], [2, 3], [2], [2]]),
        bbox_coder=dict(
            type='DeltaXYWHBBoxCoder',
            target_means=[.0, .0, .0, .0],
            target_stds=[0.1, 0.1, 0.2, 0.2])),
    # model training and testing settings
    train_cfg=dict(
        assigner=dict(
            type='MaxIoUAssigner',
            pos_iou_thr=0.5,
            neg_iou_thr=0.5,
            min_pos_iou=0.,
            ignore_iof_thr=-1,
            gt_max_assign_all=False),
        sampler=dict(type='PseudoSampler'),
        smoothl1_beta=1.,
        allowed_border=-1,
        pos_weight=-1,
        neg_pos_ratio=3,
        debug=False),
    test_cfg=dict(
        nms_pre=1000,
        nms=dict(type='nms', iou_threshold=0.45),
        min_bbox_size=0,
        score_thr=0.02,
        max_per_img=200))

#########################################################################################################
# DATASET

# dataset settings
dataset_type = 'CocoDataset'
data_root = 'data/balloon/'
backend_args = None
metainfo = {
    'classes': ('balloon', ),
    'palette': [
        (220, 20, 60),
    ]
}

train_pipeline = [
    dict(type='LoadImageFromFile', backend_args=backend_args),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(type='Resize', scale=(img_size, img_size), keep_ratio=True),
    dict(type='RandomFlip', prob=0.5),
    dict(type='PackDetInputs')
]
test_pipeline = [
    dict(type='LoadImageFromFile', backend_args=backend_args),
    dict(type='Resize', scale=(img_size, img_size), keep_ratio=True),
    # If you don't have a gt annotation, delete the pipeline
    dict(type='LoadAnnotations', with_bbox=True),
    dict(
        type='PackDetInputs',
        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
                   'scale_factor'))
]

# batch_sampler=dict(type='AspectRatioBatchSampler'),
train_dataloader = dict(
    batch_size=2,
    num_workers=2,
    persistent_workers=True,
    sampler=dict(type='DefaultSampler', shuffle=True),
    dataset=dict(
        metainfo=metainfo,
        type=dataset_type,
        data_root=data_root,
        ann_file='train.json',
        data_prefix=dict(img='train/'),
        filter_cfg=dict(filter_empty_gt=True, min_size=32),
        pipeline=train_pipeline,
        backend_args=backend_args))

val_dataloader = dict(
    batch_size=1,
    num_workers=2,
    persistent_workers=True,
    drop_last=False,
    sampler=dict(type='DefaultSampler', shuffle=False),
    dataset=dict(
        metainfo=metainfo,
        type=dataset_type,
        data_root=data_root,
        ann_file='val.json',
        data_prefix=dict(img='val/'),
        test_mode=True,
        pipeline=test_pipeline,
        backend_args=backend_args))
test_dataloader = val_dataloader

val_evaluator = dict(
    type='CocoMetric',
    ann_file=data_root + 'val.json',
    metric='bbox',
    format_only=False,
    backend_args=backend_args)

test_evaluator = val_evaluator

# inference on test dataset and
# format the output results for submission.
# test_dataloader = dict(
#     batch_size=1,
#     num_workers=2,
#     persistent_workers=True,
#     drop_last=False,
#     sampler=dict(type='DefaultSampler', shuffle=False),
#     dataset=dict(
#         type=dataset_type,
#         data_root=data_root,
#         ann_file=data_root + 'annotations/image_info_test-dev2017.json',
#         data_prefix=dict(img='test2017/'),
#         test_mode=True,
#         pipeline=test_pipeline))
# test_evaluator = dict(
#     type='CocoMetric',
#     metric='bbox',
#     format_only=True,
#     ann_file=data_root + 'annotations/image_info_test-dev2017.json',
#     outfile_prefix='./work_dirs/coco_detection/test')

#########################################################################################################
# training schedule for 1x
EPOCHS = 12
train_cfg = dict(type='EpochBasedTrainLoop', max_epochs=EPOCHS, val_interval=1)
val_cfg = dict(type='ValLoop')
test_cfg = dict(type='TestLoop')

# learning rate
param_scheduler = [
    dict(
        type='LinearLR', start_factor=0.001, by_epoch=False, begin=0, end=500),
    dict(
        type='MultiStepLR',
        begin=0,
        end=12,
        by_epoch=True,
        milestones=[8, 11],
        gamma=0.1)
]

# optimizer
optim_wrapper = dict(
    type='OptimWrapper',
    optimizer=dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001))

# Default setting for scaling LR automatically
#   - `enable` means enable scaling LR automatically
#       or not by default.
#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
auto_scale_lr = dict(enable=False, base_batch_size=16)

#########################################################################################################
default_scope = 'mmdet'

default_hooks = dict(
    timer=dict(type='IterTimerHook'),
    logger=dict(type='LoggerHook', interval=50),
    param_scheduler=dict(type='ParamSchedulerHook'),
    checkpoint=dict(type='CheckpointHook', interval=1),
    sampler_seed=dict(type='DistSamplerSeedHook'),
    visualization=dict(type='DetVisualizationHook'))

env_cfg = dict(
    cudnn_benchmark=False,
    mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0),
    dist_cfg=dict(backend='nccl'),
)

vis_backends = [dict(type='LocalVisBackend')]
visualizer = dict(
    type='DetLocalVisualizer', vis_backends=vis_backends, name='visualizer')
log_processor = dict(type='LogProcessor', window_size=50, by_epoch=True)

log_level = 'INFO'
load_from = None
resume = False

#########################################################################################################

#########################################################################################################

[Regina971]

Have you resolved this issue? Can you provide me with some details, please?

[pigfishgogogo]

did you solve the problem? i have this problem too, can you git some suggenstion？

[flywithliye]

I encountered the same issue yesterday, and upon reviewing the code, I found a temporary solution (for the main branch).

TEMPORARY FIX

1. Navigate to the code file mmdetection/mmdet/models/dense_heads/base_dense_head.py
2. At line 376, within the if-statement, comment out getattr(self.loss_cls, 'custom_cls_channels', False): and replace it with False:

Before modification

if getattr(self.loss_cls, 'custom_cls_channels', False):
    scores = self.loss_cls.get_activation(cls_score)
elif self.use_sigmoid_cls:
    scores = cls_score.sigmoid()
else:
    # remind that we set FG labels to [0, num_class-1]
    # since mmdet v2.0
    # BG cat_id: num_class
    scores = cls_score.softmax(-1)[:, :-1]

After modification

if False: # getattr(self.loss_cls, 'custom_cls_channels', False):  # Change made here
    scores = self.loss_cls.get_activation(cls_score)
elif self.use_sigmoid_cls:
    scores = cls_score.sigmoid()
else:
    # remind that we set FG labels to [0, num_class-1]
    # since mmdet v2.0
    # BG cat_id: num_class
    scores = cls_score.softmax(-1)[:, :-1]

3. By doing this, we force the 1st condition to be False, thus relying on the subsequent conditions. In earlier versions, this condition did not exist; it appears to be newly introduced. While I'm uncertain about the root cause of the malfunction, this adjustment seems to resolve the issue for now.

[pigfishgogogo]

Thank you very much for the detailed solution! -- It worked ! As a newbie, this problem has been bothering me for two days. To add: I just followed your method in the current mmdetection directory at first and it didn't work; but the bug was solved after I modified mmdet in the anaconda virtual environment!

[leileilei2000]

非常感谢您的详细解决方案！--成功了！作为一个新手，这个问题已经困扰了我两天。补充：我一开始只是在当前的mmdetection目录中遵循了您的方法，但没有用;但是在我在 anaconda 虚拟环境中修改 mmdet 后，该错误得到了解决！

very cool！！！

[SwjtuerFz]

I encountered the same issue yesterday, and upon reviewing the code, I found a temporary solution (for the main branch).

TEMPORARY FIX

1. Navigate to the code file mmdetection/mmdet/models/dense_heads/base_dense_head.py
2. At line 376, within the if-statement, comment out getattr(self.loss_cls, 'custom_cls_channels', False): and replace it with False:

Before modification

if getattr(self.loss_cls, 'custom_cls_channels', False):
    scores = self.loss_cls.get_activation(cls_score)
elif self.use_sigmoid_cls:
    scores = cls_score.sigmoid()
else:
    # remind that we set FG labels to [0, num_class-1]
    # since mmdet v2.0
    # BG cat_id: num_class
    scores = cls_score.softmax(-1)[:, :-1]

After modification

if False: # getattr(self.loss_cls, 'custom_cls_channels', False):  # Change made here
    scores = self.loss_cls.get_activation(cls_score)
elif self.use_sigmoid_cls:
    scores = cls_score.sigmoid()
else:
    # remind that we set FG labels to [0, num_class-1]
    # since mmdet v2.0
    # BG cat_id: num_class
    scores = cls_score.softmax(-1)[:, :-1]

3. By doing this, we force the 1st condition to be False, thus relying on the subsequent conditions. In earlier versions, this condition did not exist; it appears to be newly introduced. While I'm uncertain about the root cause of the malfunction, this adjustment seems to resolve the issue for now.

Thanks! It is works!

[mookiechoi]

Thank you!! I also solved this problem.