search

open-mmlab / mmdetection

OpenMMLab Detection Toolbox and Benchmark
https://mmdetection.readthedocs.io
Apache License 2.0
29.14k stars 9.39k forks source link

Multiscale test error when testing Cascade-Mask RCNN: "TypeError: list indices must be integers or slices, not tuple" #5110

Closed amobiny closed 3 years ago

amobiny commented 3 years ago

Thanks for your error report and we appreciate it a lot.

Checklist

  1. I have searched related issues but cannot get the expected help.
  2. I have read the FAQ documentation but cannot get the expected help.
  3. The bug has not been fixed in the latest version.

Describe the bug

I've successfully trained a Cascade-Mask RCNN model on my custom data. For testing, it works fine if I perform the inference on only one scale, but fails when performing the multi-scale test. The weird part is that it runs properly and breaks halfway through as you can see in the Traceback down here (not on the same image; it changes randomly every time and never reaches the end).

@yhcao6 I found https://github.com/open-mmlab/mmdetection/issues/2308 and https://github.com/open-mmlab/mmdetection/pull/2349/files which are closely related to this issue, but I think the error I'm getting is a completely different issue and happens when trying to convert the bitmap mask into RLE code in /mmdet/core/mask/utils.py

Reproduction

  1. What command or script did you run?

I ran

./tools/dist_test.sh /lhome/amobiny/config.py /lhome1/amobiny/work_dir/epoch_1.pth 4 --eval segm bbox

and here is a copy of my full config.py file

# model settings
model = dict(
    type='CascadeRCNN',
    pretrained='open-mmlab://resnext101_64x4d',
    backbone=dict(
        type='ResNeXt',
        depth=101,
        groups=64,
        base_width=4,
        num_stages=4,
        out_indices=(0, 1, 2, 3),
        frozen_stages=1,
        norm_cfg=dict(type='BN', requires_grad=True),
        norm_eval=True,
        style='pytorch'),
    neck=dict(
        type='FPN',
        in_channels=[256, 512, 1024, 2048],
        out_channels=256,
        num_outs=5),
    rpn_head=dict(
        type='RPNHead',
        in_channels=256,
        feat_channels=256,
        anchor_generator=dict(
            type='AnchorGenerator',
            scales=[8],
            ratios=[0.5, 1.0, 2.0],
            strides=[4, 8, 16, 32, 64]),
        bbox_coder=dict(
            type='DeltaXYWHBBoxCoder',
            target_means=[.0, .0, .0, .0],
            target_stds=[1.0, 1.0, 1.0, 1.0]),
        loss_cls=dict(
            type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
        loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
    roi_head=dict(
        type='CascadeRoIHead',
        num_stages=3,
        stage_loss_weights=[1, 0.5, 0.25],
        bbox_roi_extractor=dict(
            type='SingleRoIExtractor',
            roi_layer=dict(type='RoIAlign', output_size=7, sampling_ratio=0),
            out_channels=256,
            featmap_strides=[4, 8, 16, 32]),
        bbox_head=[
            dict(
                type='Shared2FCBBoxHead',
                in_channels=256,
                fc_out_channels=1024,
                roi_feat_size=7,
                num_classes=3,
                bbox_coder=dict(
                    type='DeltaXYWHBBoxCoder',
                    target_means=[0., 0., 0., 0.],
                    target_stds=[0.1, 0.1, 0.2, 0.2]),
                reg_class_agnostic=True,
                loss_cls=dict(
                    type='CrossEntropyLoss',
                    use_sigmoid=False,
                    loss_weight=1.0),
                loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
                               loss_weight=1.0)),
            dict(
                type='Shared2FCBBoxHead',
                in_channels=256,
                fc_out_channels=1024,
                roi_feat_size=7,
                num_classes=3,
                bbox_coder=dict(
                    type='DeltaXYWHBBoxCoder',
                    target_means=[0., 0., 0., 0.],
                    target_stds=[0.05, 0.05, 0.1, 0.1]),
                reg_class_agnostic=True,
                loss_cls=dict(
                    type='CrossEntropyLoss',
                    use_sigmoid=False,
                    loss_weight=1.0),
                loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
                               loss_weight=1.0)),
            dict(
                type='Shared2FCBBoxHead',
                in_channels=256,
                fc_out_channels=1024,
                roi_feat_size=7,
                num_classes=3,
                bbox_coder=dict(
                    type='DeltaXYWHBBoxCoder',
                    target_means=[0., 0., 0., 0.],
                    target_stds=[0.033, 0.033, 0.067, 0.067]),
                reg_class_agnostic=True,
                loss_cls=dict(
                    type='CrossEntropyLoss',
                    use_sigmoid=False,
                    loss_weight=1.0),
                loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
        ],
        mask_roi_extractor=dict(
            type='SingleRoIExtractor',
            roi_layer=dict(type='RoIAlign', output_size=14, sampling_ratio=0),
            out_channels=256,
            featmap_strides=[4, 8, 16, 32]),
        mask_head=dict(
            type='FCNMaskHead',
            num_convs=4,
            in_channels=256,
            conv_out_channels=256,
            num_classes=3,
            loss_mask=dict(
                type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))),
    # model training and testing settings
    train_cfg=dict(
        rpn=dict(
            assigner=dict(
                type='MaxIoUAssigner',
                pos_iou_thr=0.7,
                neg_iou_thr=0.3,
                min_pos_iou=0.3,
                match_low_quality=True,
                ignore_iof_thr=-1),
            sampler=dict(
                type='RandomSampler',
                num=256,
                pos_fraction=0.5,
                neg_pos_ub=-1,
                add_gt_as_proposals=False),
            allowed_border=0,
            pos_weight=-1,
            debug=False),
        rpn_proposal=dict(
            nms_pre=2000,
            max_per_img=2000,
            nms=dict(type='nms', iou_threshold=0.7),
            min_bbox_size=0),
        rcnn=[
            dict(
                assigner=dict(
                    type='MaxIoUAssigner',
                    pos_iou_thr=0.5,
                    neg_iou_thr=0.5,
                    min_pos_iou=0.5,
                    match_low_quality=False,
                    ignore_iof_thr=-1),
                sampler=dict(
                    type='RandomSampler',
                    num=512,
                    pos_fraction=0.25,
                    neg_pos_ub=-1,
                    add_gt_as_proposals=True),
                mask_size=28,
                pos_weight=-1,
                debug=False),
            dict(
                assigner=dict(
                    type='MaxIoUAssigner',
                    pos_iou_thr=0.6,
                    neg_iou_thr=0.6,
                    min_pos_iou=0.6,
                    match_low_quality=False,
                    ignore_iof_thr=-1),
                sampler=dict(
                    type='RandomSampler',
                    num=512,
                    pos_fraction=0.25,
                    neg_pos_ub=-1,
                    add_gt_as_proposals=True),
                mask_size=28,
                pos_weight=-1,
                debug=False),
            dict(
                assigner=dict(
                    type='MaxIoUAssigner',
                    pos_iou_thr=0.7,
                    neg_iou_thr=0.7,
                    min_pos_iou=0.7,
                    match_low_quality=False,
                    ignore_iof_thr=-1),
                sampler=dict(
                    type='RandomSampler',
                    num=512,
                    pos_fraction=0.25,
                    neg_pos_ub=-1,
                    add_gt_as_proposals=True),
                mask_size=28,
                pos_weight=-1,
                debug=False)
        ]),
    test_cfg=dict(
        rpn=dict(
            nms_pre=1000,
            max_per_img=1000,
            nms=dict(type='nms', iou_threshold=0.7),
            min_bbox_size=0),
        rcnn=dict(
            score_thr=0.05,
            nms=dict(type='nms', iou_threshold=0.5),
            max_per_img=100,
            mask_thr_binary=0.5)))

# dataset settings
dataset_type = 'CocoDataset'
data_root = '/lhome1/amobiny/datasets/Pole_1.2_SC5_SRM_rotated/'
img_norm_cfg = dict(
    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
    dict(type='Resize',
         img_scale=[(1024, 992), (1024, 512)],     # original was (992, 1024)
         multiscale_mode='range',
         keep_ratio=True),
    dict(type='RandomFlip', flip_ratio=0.5),
    dict(type='Normalize', **img_norm_cfg),
    dict(type='Pad', size_divisor=32),
    dict(type='DefaultFormatBundle'),
    dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
]
test_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        type='MultiScaleFlipAug',
        img_scale=[(1024, 992), (1024, 752), (1024, 512)],
        # img_scale=(1024, 992),
        flip=False,
        transforms=[
            dict(type='Resize', keep_ratio=True),
            dict(type='RandomFlip', flip_ratio=0.0),
            dict(type='Normalize', **img_norm_cfg),
            dict(type='Pad', size_divisor=32),
            dict(type='ImageToTensor', keys=['img']),
            dict(type='Collect', keys=['img']),
        ])
]
data = dict(
    samples_per_gpu=1,
    workers_per_gpu=1,
    train=dict(
        type=dataset_type,
        ann_file=data_root + 'annotations/coco_pole_srm_1.2_validation.json',
        img_prefix=data_root + 'images/',
        classes=['pole', 'delineator_post', 'inverse_U'],
        pipeline=train_pipeline),
    val=dict(
        type=dataset_type,
        ann_file=data_root + 'annotations/coco_pole_srm_1.2_validation.json',
        # ann_file=data_root + 'annotations/coco_arrow_subset_lrm_sc5_validation.json',
        img_prefix=data_root + 'images/',
        classes=['pole', 'delineator_post', 'inverse_U'],
        pipeline=test_pipeline),
    test=dict(
        type=dataset_type,
        ann_file=data_root + 'annotations/coco_pole_srm_1.2_validation.json',
        # ann_file=data_root + 'annotations/coco_arrow_subset_lrm_sc5_validation.json',
        img_prefix=data_root + 'images/',
        classes=['pole', 'delineator_post', 'inverse_U'],
        pipeline=test_pipeline))
evaluation = dict(metric=['bbox', 'segm'])

# optimizer
optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
optimizer_config = dict(grad_clip=None)
# learning policy
lr_config = dict(
    policy='step',
    warmup='linear',
    warmup_iters=500,
    warmup_ratio=0.001,
    step=[16, 19])
runner = dict(type='EpochBasedRunner', max_epochs=20)

checkpoint_config = dict(interval=1)
# yapf:disable
log_config = dict(
    interval=50,
    hooks=[
        dict(type='TextLoggerHook'),
        # dict(type='TensorboardLoggerHook')
    ])
# yapf:enable
custom_hooks = [dict(type='NumClassCheckHook')]
dist_params = dict(backend='nccl')
log_level = 'INFO'
work_dir = './work_dirs/run3_mmdet2'
load_from = None
resume_from = None
workflow = [('train', 1)]
  1. Did you make any modifications on the code or config? Did you understand what you have modified?

I've made no major change to the default Cascade-Mask RCNN model.

  1. What dataset did you use?

Environment

  1. Please run python mmdet/utils/collect_env.py to collect necessary environment information and paste it here.

fatal: Not a git repository (or any parent up to mount point /lhome1) Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set). sys.platform: linux Python: 3.7.10 (default, Feb 26 2021, 18:47:35) [GCC 7.3.0] CUDA available: True GPU 0,1,2,3: Quadro RTX 8000 CUDA_HOME: /usr/local/cuda-10.1 NVCC: Cuda compilation tools, release 10.1, V10.1.243 GCC: gcc (Ubuntu 5.4.0-6ubuntu1~16.04.12) 5.4.0 20160609 PyTorch: 1.7.0 PyTorch compiling details: PyTorch built with:

TorchVision: 0.8.0 OpenCV: 4.5.1 MMCV: 1.3.3 MMCV Compiler: GCC 7.3 MMCV CUDA Compiler: 10.1 MMDetection: 2.11.0+

  1. You may add addition that may be helpful for locating the problem, such as
    • How you installed PyTorch [e.g., pip, conda, source]
    • Other environment variables that may be related (such as $PATH, $LD_LIBRARY_PATH, $PYTHONPATH, etc.)

Error traceback If applicable, paste the error trackback here.

[>>>>>>>>>>>>>>>>>>>>>>>>>                         ] 5056/10009, 10.5 task/s, elapsed: 481s, ETA:   471sTraceback (most recent call last):
  File "/dump/algossd/ary/mmdetection2/tools/test.py", line 220, in <module>
    main()
  File "/dump/algossd/ary/mmdetection2/tools/test.py", line 197, in main
    args.gpu_collect)
  File "/dump/algossd/ary/mmdetection2/mmdet/apis/test.py", line 101, in multi_gpu_test
    for bbox_results, mask_results in result]
  File "/dump/algossd/ary/mmdetection2/mmdet/apis/test.py", line 101, in <listcomp>
    for bbox_results, mask_results in result]
  File "/dump/algossd/ary/mmdetection2/mmdet/core/mask/utils.py", line 58, in encode_mask_results
    cls_segm[:, :, np.newaxis], order='F',
TypeError: list indices must be integers or slices, not tuple
Traceback (most recent call last):
  File "/dump/algossd/ary/anaconda3/envs/mmlab2/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/dump/algossd/ary/anaconda3/envs/mmlab2/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/dump/algossd/ary/anaconda3/envs/mmlab2/lib/python3.7/site-packages/torch/distributed/launch.py", line 260, in <module>
    main()
  File "/dump/algossd/ary/anaconda3/envs/mmlab2/lib/python3.7/site-packages/torch/distributed/launch.py", line 256, in main
    cmd=cmd)
subprocess.CalledProcessError: Command '['/dump/algossd/ary/anaconda3/envs/mmlab2/bin/python', '-u', '/dump/algossd/ary/mmdetection2/tools/test.py', '--local_rank=3', '/lhome1/amobiny/workspace/Pole_1.2_SC5_SRM/config_mmdet2.py', '/lhome1/amobiny/workspace/Pole_1.2_SC5_SRM/work_dirs/run3_mmdet2/epoch_1.pth', '--launcher', 'pytorch', '--eval', 'segm', 'bbox']' returned non-zero exit status 1.
hhaAndroid commented 3 years ago

@amobiny Hi, Did you modify the code? Can you test it with a single machine first? Thank you!

amobiny commented 3 years ago

Hi @hhaAndroid The same error pops up even when running on 1 gpu. I was able to find the problem and resolve it. The reason is that sometimes when there is no object detected in an image, the mask_results in /mmdet/apis/test (i both single_gpu_test and multi_gpu_test) becomes [[[], [], []]] (which is a list of length 1), in a 3-class problem, for example. However, it should be [[], [], []] which is a list of length 3! This causes the error in /mmdet/core/mask/utils.py in the encode_mask_results function. I modified the encode_mask_results code and it works all good now.

hhaAndroid commented 3 years ago

Hi @hhaAndroid The same error pops up even when running on 1 gpu. I was able to find the problem and resolve it. The reason is that sometimes when there is no object detected in an image, the mask_results in /mmdet/apis/test (i both single_gpu_test and multi_gpu_test) becomes [[[], [], []]] (which is a list of length 1), in a 3-class problem, for example. However, it should be [[], [], []] which is a list of length 3! This causes the error in /mmdet/core/mask/utils.py in the encode_mask_results function. I modified the encode_mask_results code and it works all good now.

OK, Thank you, I will check it and fix.