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

use rtmdet COCO config file in wider_face datasets map=0 #11337

Open keeper-jie opened 8 months ago

keeper-jie commented 8 months ago
  1. when i use rtmdet coco config in wider_face dataset, the 80 epoch map is 0.3, but in 90 epoch map is 0, I do not know how to find the reason, pleaser help me
  2. enviroment is Nvidia A6000 GPU, CUDA 11.8, mmdetection dcoker with 3.2.0 version
  3. config file as bleow: 
    auto_scale_lr = dict(base_batch_size=16, enable=True)
backend_args = None
base_lr = 0.004
checkpoint = 'https://download.openmmlab.com/mmdetection/v3.0/rtmdet/cspnext_rsb_pretrain/cspnext-s_imagenet_600e.pth'
custom_hooks = [
dict(
    ema_type='ExpMomentumEMA',
    momentum=0.0002,
    priority=49,
    type='EMAHook',
    update_buffers=True),
dict(
    switch_epoch=280,
    switch_pipeline=[
        dict(backend_args=None, type='LoadImageFromFile'),
        dict(type='LoadAnnotations', with_bbox=True),
        dict(
            keep_ratio=True,
            ratio_range=(
                0.5,
                2.0,
            ),
            scale=(
                640,
                640,
            ),
            type='RandomResize'),
        dict(crop_size=(
            640,
            640,
        ), type='RandomCrop'),
        dict(type='YOLOXHSVRandomAug'),
        dict(prob=0.5, type='RandomFlip'),
        dict(
            pad_val=dict(img=(
                114,
                114,
                114,
            )),
            size=(
                640,
                640,
            ),
            type='Pad'),
        dict(type='PackDetInputs'),
    ],
    type='PipelineSwitchHook'),
]
data_root = '/mmdetection/data/widerface/'
dataset_type = 'CocoDataset'
default_hooks = dict(
checkpoint=dict(interval=10, max_keep_ckpts=3, type='CheckpointHook'),
logger=dict(interval=50, type='LoggerHook'),
param_scheduler=dict(type='ParamSchedulerHook'),
sampler_seed=dict(type='DistSamplerSeedHook'),
timer=dict(type='IterTimerHook'),
visualization=dict(type='DetVisualizationHook'))
default_scope = 'mmdet'
env_cfg = dict(
cudnn_benchmark=False,
dist_cfg=dict(backend='nccl'),
mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0))
img_scales = [
(
    640,
    640,
),
(
    320,
    320,
),
(
    960,
    960,
),
]
interval = 10
launcher = 'none'
load_from = './checkpoints/rtmdet_s_8xb32-300e_coco_20220905_161602-387a891e.pth'
log_level = 'INFO'
log_processor = dict(by_epoch=True, type='LogProcessor', window_size=50)
max_epochs = 300
metainfo = dict(
classes=('face', ), palette=[
    (
        220,
        20,
        60,
    ),
])
model = dict(
backbone=dict(
    act_cfg=dict(inplace=True, type='SiLU'),
    arch='P5',
    channel_attention=True,
    deepen_factor=0.33,
    expand_ratio=0.5,
    init_cfg=dict(
        checkpoint=
        'https://download.openmmlab.com/mmdetection/v3.0/rtmdet/cspnext_rsb_pretrain/cspnext-s_imagenet_600e.pth',
        prefix='backbone.',
        type='Pretrained'),
    norm_cfg=dict(type='SyncBN'),
    type='CSPNeXt',
    widen_factor=0.5),
bbox_head=dict(
    act_cfg=dict(inplace=True, type='SiLU'),
    anchor_generator=dict(
        offset=0, strides=[
            8,
            16,
            32,
        ], type='MlvlPointGenerator'),
    bbox_coder=dict(type='DistancePointBBoxCoder'),
    exp_on_reg=False,
    feat_channels=128,
    in_channels=128,
    loss_bbox=dict(loss_weight=2.0, type='GIoULoss'),
    loss_cls=dict(
        beta=2.0,
        loss_weight=1.0,
        type='QualityFocalLoss',
        use_sigmoid=True),
    norm_cfg=dict(type='SyncBN'),
    num_classes=1,
    pred_kernel_size=1,
    share_conv=True,
    stacked_convs=2,
    type='RTMDetSepBNHead',
    with_objectness=False),
data_preprocessor=dict(
    batch_augments=None,
    bgr_to_rgb=False,
    mean=[
        103.53,
        116.28,
        123.675,
    ],
    std=[
        57.375,
        57.12,
        58.395,
    ],
    type='DetDataPreprocessor'),
neck=dict(
    act_cfg=dict(inplace=True, type='SiLU'),
    expand_ratio=0.5,
    in_channels=[
        128,
        256,
        512,
    ],
    norm_cfg=dict(type='SyncBN'),
    num_csp_blocks=1,
    out_channels=128,
    type='CSPNeXtPAFPN'),
test_cfg=dict(
    max_per_img=300,
    min_bbox_size=0,
    nms=dict(iou_threshold=0.65, type='nms'),
    nms_pre=30000,
    score_thr=0.001),
train_cfg=dict(
    allowed_border=-1,
    assigner=dict(topk=13, type='DynamicSoftLabelAssigner'),
    debug=False,
    pos_weight=-1),
type='RTMDet')
optim_wrapper = dict(
optimizer=dict(lr=0.004, type='AdamW', weight_decay=0.05),
paramwise_cfg=dict(
    bias_decay_mult=0, bypass_duplicate=True, norm_decay_mult=0),
type='OptimWrapper')
param_scheduler = [
dict(
    begin=0, by_epoch=False, end=1000, start_factor=1e-05,
    type='LinearLR'),
dict(
    T_max=150,
    begin=150,
    by_epoch=True,
    convert_to_iter_based=True,
    end=300,
    eta_min=0.0002,
    type='CosineAnnealingLR'),
]
resume = False
stage2_num_epochs = 20
test_cfg = dict(type='TestLoop')
test_dataloader = dict(
batch_size=5,
dataset=dict(
    ann_file='WIDER_val/wider_face_val_annot_coco_style.json',
    backend_args=None,
    data_prefix=dict(img='WIDER_val/images'),
    data_root='/mmdetection/data/widerface/',
    metainfo=dict(classes=('face', ), palette=[
        (
            220,
            20,
            60,
        ),
    ]),
    pipeline=[
        dict(backend_args=None, type='LoadImageFromFile'),
        dict(keep_ratio=True, scale=(
            640,
            640,
        ), type='Resize'),
        dict(
            pad_val=dict(img=(
                114,
                114,
                114,
            )),
            size=(
                640,
                640,
            ),
            type='Pad'),
        dict(type='LoadAnnotations', with_bbox=True),
        dict(
            meta_keys=(
                'img_id',
                'img_path',
                'ori_shape',
                'img_shape',
                'scale_factor',
            ),
            type='PackDetInputs'),
    ],
    test_mode=True,
    type='CocoDataset'),
drop_last=False,
num_workers=10,
persistent_workers=True,
sampler=dict(shuffle=False, type='DefaultSampler'))
test_evaluator = dict(
ann_file=
'/mmdetection/data/widerface/WIDER_val/wider_face_val_annot_coco_style.json',
backend_args=None,
format_only=False,
metric='bbox',
proposal_nums=(
    100,
    1,
    10,
),
type='CocoMetric')
test_pipeline = [
dict(backend_args=None, type='LoadImageFromFile'),
dict(keep_ratio=True, scale=(
    640,
    640,
), type='Resize'),
dict(pad_val=dict(img=(
    114,
    114,
    114,
)), size=(
    640,
    640,
), type='Pad'),
dict(type='LoadAnnotations', with_bbox=True),
dict(
    meta_keys=(
        'img_id',
        'img_path',
        'ori_shape',
        'img_shape',
        'scale_factor',
    ),
    type='PackDetInputs'),
]
train_batch_size_per_gpu = 96
train_cfg = dict(
dynamic_intervals=[
    (
        280,
        1,
    ),
],
max_epochs=300,
type='EpochBasedTrainLoop',
val_interval=10)
train_dataloader = dict(
batch_sampler=None,
batch_size=96,
dataset=dict(
    ann_file='WIDER_train/wider_face_train_annot_coco_style.json',
    backend_args=None,
    data_prefix=dict(img='WIDER_train/images'),
    data_root='/mmdetection/data/widerface/',
    filter_cfg=dict(filter_empty_gt=True, min_size=32),
    metainfo=dict(classes=('face', ), palette=[
        (
            220,
            20,
            60,
        ),
    ]),
    pipeline=[
        dict(backend_args=None, type='LoadImageFromFile'),
        dict(type='LoadAnnotations', with_bbox=True),
        dict(img_scale=(
            640,
            640,
        ), pad_val=114.0, type='CachedMosaic'),
        dict(
            keep_ratio=True,
            ratio_range=(
                0.5,
                2.0,
            ),
            scale=(
                1280,
                1280,
            ),
            type='RandomResize'),
        dict(crop_size=(
            640,
            640,
        ), type='RandomCrop'),
        dict(type='YOLOXHSVRandomAug'),
        dict(prob=0.5, type='RandomFlip'),
        dict(
            pad_val=dict(img=(
                114,
                114,
                114,
            )),
            size=(
                640,
                640,
            ),
            type='Pad'),
        dict(
            img_scale=(
                640,
                640,
            ),
            max_cached_images=20,
            pad_val=(
                114,
                114,
                114,
            ),
            ratio_range=(
                1.0,
                1.0,
            ),
            type='CachedMixUp'),
        dict(type='PackDetInputs'),
    ],
    type='CocoDataset'),
num_workers=8,
persistent_workers=True,
pin_memory=True,
sampler=dict(shuffle=True, type='DefaultSampler'))
train_num_workers = 8
train_pipeline = [
dict(backend_args=None, type='LoadImageFromFile'),
dict(type='LoadAnnotations', with_bbox=True),
dict(img_scale=(
    640,
    640,
), pad_val=114.0, type='CachedMosaic'),
dict(
    keep_ratio=True,
    ratio_range=(
        0.5,
        2.0,
    ),
    scale=(
        1280,
        1280,
    ),
    type='RandomResize'),
dict(crop_size=(
    640,
    640,
), type='RandomCrop'),
dict(type='YOLOXHSVRandomAug'),
dict(prob=0.5, type='RandomFlip'),
dict(pad_val=dict(img=(
    114,
    114,
    114,
)), size=(
    640,
    640,
), type='Pad'),
dict(
    img_scale=(
        640,
        640,
    ),
    max_cached_images=20,
    pad_val=(
        114,
        114,
        114,
    ),
    ratio_range=(
        1.0,
        1.0,
    ),
    type='CachedMixUp'),
dict(type='PackDetInputs'),
]
train_pipeline_stage2 = [
dict(backend_args=None, type='LoadImageFromFile'),
dict(type='LoadAnnotations', with_bbox=True),
dict(
    keep_ratio=True,
    ratio_range=(
        0.5,
        2.0,
    ),
    scale=(
        640,
        640,
    ),
    type='RandomResize'),
dict(crop_size=(
    640,
    640,
), type='RandomCrop'),
dict(type='YOLOXHSVRandomAug'),
dict(prob=0.5, type='RandomFlip'),
dict(pad_val=dict(img=(
    114,
    114,
    114,
)), size=(
    640,
    640,
), type='Pad'),
dict(type='PackDetInputs'),
]
tta_model = dict(
tta_cfg=dict(max_per_img=100, nms=dict(iou_threshold=0.6, type='nms')),
type='DetTTAModel')
tta_pipeline = [
dict(backend_args=None, type='LoadImageFromFile'),
dict(
    transforms=[
        [
            dict(keep_ratio=True, scale=(
                640,
                640,
            ), type='Resize'),
            dict(keep_ratio=True, scale=(
                320,
                320,
            ), type='Resize'),
            dict(keep_ratio=True, scale=(
                960,
                960,
            ), type='Resize'),
        ],
        [
            dict(prob=1.0, type='RandomFlip'),
            dict(prob=0.0, type='RandomFlip'),
        ],
        [
            dict(
                pad_val=dict(img=(
                    114,
                    114,
                    114,
                )),
                size=(
                    960,
                    960,
                ),
                type='Pad'),
        ],
        [
            dict(type='LoadAnnotations', with_bbox=True),
        ],
        [
            dict(
                meta_keys=(
                    'img_id',
                    'img_path',
                    'ori_shape',
                    'img_shape',
                    'scale_factor',
                    'flip',
                    'flip_direction',
                ),
                type='PackDetInputs'),
        ],
    ],
    type='TestTimeAug'),
]
val_cfg = dict(type='ValLoop')
val_dataloader = dict(
batch_size=5,
dataset=dict(
    ann_file='WIDER_val/wider_face_val_annot_coco_style.json',
    backend_args=None,
    data_prefix=dict(img='WIDER_val/images'),
    data_root='/mmdetection/data/widerface/',
    metainfo=dict(classes=('face', ), palette=[
        (
            220,
            20,
            60,
        ),
    ]),
    pipeline=[
        dict(backend_args=None, type='LoadImageFromFile'),
        dict(keep_ratio=True, scale=(
            640,
            640,
        ), type='Resize'),
        dict(
            pad_val=dict(img=(
                114,
                114,
                114,
            )),
            size=(
                640,
                640,
            ),
            type='Pad'),
        dict(type='LoadAnnotations', with_bbox=True),
        dict(
            meta_keys=(
                'img_id',
                'img_path',
                'ori_shape',
                'img_shape',
                'scale_factor',
            ),
            type='PackDetInputs'),
    ],
    test_mode=True,
    type='CocoDataset'),
drop_last=False,
num_workers=10,
persistent_workers=True,
sampler=dict(shuffle=False, type='DefaultSampler'))
val_evaluator = dict(
ann_file=
'/mmdetection/data/widerface/WIDER_val/wider_face_val_annot_coco_style.json',
backend_args=None,
format_only=False,
metric='bbox',
proposal_nums=(
    100,
    1,
    10,
),
type='CocoMetric')
vis_backends = [
dict(type='LocalVisBackend'),
]
visualizer = dict(
name='visualizer',
type='DetLocalVisualizer',
vis_backends=[
    dict(type='LocalVisBackend'),
])
work_dir = './work_dirs/rtmdet_s_1xb96-300e_face'
  4. run program command
    CUDA_VISIBLE_DEVICES=1 python ../tools/train.py ../configs/rtmdet/rtmdet_s_1xb96-300e_face.py --auto-scale-lr
  5. output as blow: 
    12/29 17:47:58 - mmengine - INFO - Epoch(val) [80][645/645]    coco/bbox_mAP: 0.3050  coco/bbox_mAP_50: 0.5480  coco/bbox_mAP_75: 0.3060  coco/bbox_mAP_s: 0.1830  coco/bbox_mAP_m: 0.5930  coco/bbox_mAP_l: 0.6830  data_time: 0.0074  time: 0.0390
12/29 17:49:28 - mmengine - INFO - Epoch(train)  [81][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:49:01  time: 1.7882  data_time: 0.1368  memory: 42078  loss: 0.7641  loss_cls: 0.3325  loss_bbox: 0.4315
12/29 17:50:49 - mmengine - INFO - Epoch(train)  [81][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:47:29  time: 1.6186  data_time: 0.0037  memory: 41906  loss: 0.7815  loss_cls: 0.3464  loss_bbox: 0.4351
12/29 17:51:43 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 17:53:12 - mmengine - INFO - Epoch(train)  [82][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:45:09  time: 1.7788  data_time: 0.1264  memory: 42168  loss: 0.7743  loss_cls: 0.3411  loss_bbox: 0.4332
12/29 17:53:37 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 17:54:35 - mmengine - INFO - Epoch(train)  [82][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:43:43  time: 1.6570  data_time: 0.0037  memory: 42144  loss: 0.7612  loss_cls: 0.3292  loss_bbox: 0.4319
12/29 17:55:30 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 17:57:00 - mmengine - INFO - Epoch(train)  [83][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:41:27  time: 1.8053  data_time: 0.1507  memory: 41823  loss: 0.7800  loss_cls: 0.3465  loss_bbox: 0.4335
12/29 17:58:23 - mmengine - INFO - Epoch(train)  [83][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:40:00  time: 1.6506  data_time: 0.0036  memory: 41814  loss: 0.7794  loss_cls: 0.3383  loss_bbox: 0.4410
12/29 17:59:18 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:00:51 - mmengine - INFO - Epoch(train)  [84][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:37:52  time: 1.8615  data_time: 0.1427  memory: 42267  loss: 0.7558  loss_cls: 0.3254  loss_bbox: 0.4305
12/29 18:02:15 - mmengine - INFO - Epoch(train)  [84][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:36:28  time: 1.6744  data_time: 0.0048  memory: 42008  loss: 0.7609  loss_cls: 0.3316  loss_bbox: 0.4293
12/29 18:03:10 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:04:41 - mmengine - INFO - Epoch(train)  [85][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:34:15  time: 1.8227  data_time: 0.1384  memory: 42171  loss: 0.7903  loss_cls: 0.3466  loss_bbox: 0.4437
12/29 18:06:01 - mmengine - INFO - Epoch(train)  [85][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:32:40  time: 1.5903  data_time: 0.0036  memory: 41711  loss: 0.7744  loss_cls: 0.3408  loss_bbox: 0.4336
12/29 18:06:54 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:08:25 - mmengine - INFO - Epoch(train)  [86][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:30:20  time: 1.8100  data_time: 0.1216  memory: 42291  loss: 0.7712  loss_cls: 0.3343  loss_bbox: 0.4369
12/29 18:09:46 - mmengine - INFO - Epoch(train)  [86][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:28:51  time: 1.6377  data_time: 0.0040  memory: 41881  loss: 0.7795  loss_cls: 0.3451  loss_bbox: 0.4343
12/29 18:10:41 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:12:11 - mmengine - INFO - Epoch(train)  [87][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:26:34  time: 1.8064  data_time: 0.1444  memory: 42257  loss: 0.7773  loss_cls: 0.3383  loss_bbox: 0.4391
12/29 18:13:34 - mmengine - INFO - Epoch(train)  [87][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:25:08  time: 1.6579  data_time: 0.0037  memory: 42181  loss: 0.7682  loss_cls: 0.3333  loss_bbox: 0.4349
12/29 18:14:30 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:15:59 - mmengine - INFO - Epoch(train)  [88][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:22:51  time: 1.7888  data_time: 0.1217  memory: 41883  loss: 0.7717  loss_cls: 0.3321  loss_bbox: 0.4395
12/29 18:17:23 - mmengine - INFO - Epoch(train)  [88][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:21:25  time: 1.6672  data_time: 0.0042  memory: 41819  loss: 0.7583  loss_cls: 0.3302  loss_bbox: 0.4281
12/29 18:18:17 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:19:45 - mmengine - INFO - Epoch(train)  [89][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:19:03  time: 1.7663  data_time: 0.1200  memory: 42185  loss: 0.7939  loss_cls: 0.3476  loss_bbox: 0.4462
12/29 18:21:08 - mmengine - INFO - Epoch(train)  [89][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:17:37  time: 1.6618  data_time: 0.0036  memory: 41868  loss: 0.7662  loss_cls: 0.3335  loss_bbox: 0.4327
12/29 18:21:41 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:22:03 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:23:33 - mmengine - INFO - Epoch(train)  [90][ 50/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:15:19  time: 1.7940  data_time: 0.1402  memory: 41891  loss: 2.7779  loss_cls: 1.6989  loss_bbox: 1.0790
12/29 18:24:56 - mmengine - INFO - Epoch(train)  [90][100/135]  base_lr: 4.0000e-03 lr: 2.4000e-02  eta: 13:13:53  time: 1.6645  data_time: 0.0037  memory: 42341  loss: 2.6869  loss_cls: 1.4744  loss_bbox: 1.2125
12/29 18:25:50 - mmengine - INFO - Exp name: rtmdet_s_1xb96-300e_face_20231229_122721
12/29 18:25:50 - mmengine - INFO - Saving checkpoint at 90 epochs
12/29 18:25:54 - mmengine - INFO - Epoch(val)  [90][ 50/645]    eta: 0:00:22  time: 0.0380  data_time: 0.0100  memory: 41754  
12/29 18:25:55 - mmengine - INFO - Epoch(val)  [90][100/645]    eta: 0:00:19  time: 0.0330  data_time: 0.0058  memory: 414  
12/29 18:25:57 - mmengine - INFO - Epoch(val)  [90][150/645]    eta: 0:00:17  time: 0.0335  data_time: 0.0063  memory: 414  
12/29 18:25:59 - mmengine - INFO - Epoch(val)  [90][200/645]    eta: 0:00:15  time: 0.0336  data_time: 0.0064  memory: 414  
12/29 18:26:00 - mmengine - INFO - Epoch(val)  [90][250/645]    eta: 0:00:13  time: 0.0333  data_time: 0.0059  memory: 414  
12/29 18:26:02 - mmengine - INFO - Epoch(val)  [90][300/645]    eta: 0:00:11  time: 0.0330  data_time: 0.0057  memory: 414  
12/29 18:26:04 - mmengine - INFO - Epoch(val)  [90][350/645]    eta: 0:00:09  time: 0.0329  data_time: 0.0057  memory: 414  
12/29 18:26:05 - mmengine - INFO - Epoch(val)  [90][400/645]    eta: 0:00:08  time: 0.0330  data_time: 0.0057  memory: 414  
12/29 18:26:07 - mmengine - INFO - Epoch(val)  [90][450/645]    eta: 0:00:06  time: 0.0332  data_time: 0.0058  memory: 414  
12/29 18:26:09 - mmengine - INFO - Epoch(val)  [90][500/645]    eta: 0:00:04  time: 0.0331  data_time: 0.0058  memory: 414  
12/29 18:26:10 - mmengine - INFO - Epoch(val)  [90][550/645]    eta: 0:00:03  time: 0.0332  data_time: 0.0060  memory: 414  
12/29 18:26:12 - mmengine - INFO - Epoch(val)  [90][600/645]    eta: 0:00:01  time: 0.0333  data_time: 0.0060  memory: 414  
12/29 18:26:31 - mmengine - INFO - Evaluating bbox...
Loading and preparing results...
DONE (t=2.56s)
creating index...
index created!
Running per image evaluation...
Evaluate annotation type *bbox*
DONE (t=85.99s).
Accumulating evaluation results...
DONE (t=3.30s).
Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.000
Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.000
Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.000
Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000
Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.000
Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.000
Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.000
Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.000
Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.000
Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000
Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.000
Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.007
12/29 18:28:08 - mmengine - INFO - bbox_mAP_copypaste: 0.000 0.000 0.000 0.000 0.000 0.000
12/29 18:28:08 - mmengine - INFO - Epoch(val) [90][645/645]    coco/bbox_mAP: 0.0000  coco/bbox_mAP_50: 0.0000  coco/bbox_mAP_75: 0.0000  coco/bbox_mAP_s: 0.0000  coco/bbox_mAP_m: 0.0000  coco/bbox_mAP_l: 0.0000  data_time: 0.0063  time: 0.0336
  6. logfile:
    20231229_122721.log
huashuzi commented 8 months ago

你好！可以向您请教一些关于rtmdet的问题吗？

keeper-jie commented 6 months ago

你好！可以向您请教一些关于rtmdet的问题吗？

可以的，Q747055484