search

open-mmlab / mmdetection

OpenMMLab Detection Toolbox and Benchmark
https://mmdetection.readthedocs.io
Apache License 2.0
29.01k stars 9.36k forks source link

Apex AMP erroring out on checkpointing after epoch and pre-trained models can't be loaded for fine-tuning/warm restart #6219

Open neel04 opened 2 years ago

neel04 commented 2 years ago

Hi, I have created a WandB report that is attached to the issue that outlines the exact model and loss logs of the model. WB_report.pdf

The model I am using is Swin transfomer with Cascade RCNN. From what I deduce, it seems that one of my checkpoints is not being loaded.

  1. It seems that the backbones checkpoint is not being loaded - but the detector can be easily loaded.

  2. Additionally, I have a dataset of about 5K images - divided into 3 classes. The Batch Size = 8; It may be overfitting, maybe not. I can't diagnose the problems 😢

  3. I might also be interpreting the model checkpoints incorrectly; I think that the checkpoints here should be for the whole model?

Some errors:-

Interestingly enough,

2021-10-01 22:08:13,462 - mmdet - INFO - Set random seed to 42, deterministic: True
2021-10-01 22:08:14,210 - mmdet - INFO - load model from: /kaggle/working/checkpoints/cascade_mask_rcnn_swin_small_patch4_window7.pth
Traceback (most recent call last):
  File "/opt/conda/lib/python3.7/site-packages/mmcv/utils/registry.py", line 52, in build_from_cfg
    return obj_cls(**args)
  File "/kaggle/working/Swin-Transformer-Object-Detection/mmdet/models/detectors/cascade_rcnn.py", line 25, in __init__
    pretrained=pretrained)
  File "/kaggle/working/Swin-Transformer-Object-Detection/mmdet/models/detectors/two_stage.py", line 48, in __init__
    self.init_weights(pretrained=pretrained)
  File "/kaggle/working/Swin-Transformer-Object-Detection/mmdet/models/detectors/two_stage.py", line 68, in init_weights
    self.backbone.init_weights(pretrained=pretrained)
  File "/kaggle/working/Swin-Transformer-Object-Detection/mmdet/models/backbones/swin_transformer.py", line 594, in init_weights
    load_checkpoint(self, pretrained, strict=False, logger=logger)
  File "/kaggle/working/Swin-Transformer-Object-Detection/mmcv_custom/checkpoint.py", line 340, in load_checkpoint
    table_current = model.state_dict()[table_key]
KeyError: 'backbone.layers.0.blocks.0.attn.relative_position_bias_table'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "tools/train.py", line 187, in <module>
    main()
  File "tools/train.py", line 161, in main
    test_cfg=cfg.get('test_cfg'))
  File "/kaggle/working/Swin-Transformer-Object-Detection/mmdet/models/builder.py", line 77, in build_detector
    return build(cfg, DETECTORS, dict(train_cfg=train_cfg, test_cfg=test_cfg))
  File "/kaggle/working/Swin-Transformer-Object-Detection/mmdet/models/builder.py", line 34, in build
    return build_from_cfg(cfg, registry, default_args)
  File "/opt/conda/lib/python3.7/site-packages/mmcv/utils/registry.py", line 55, in build_from_cfg
    raise type(e)(f'{obj_cls.__name__}: {e}')
KeyError: "CascadeRCNN: 'backbone.layers.0.blocks.0.attn.relative_position_bias_table'"

it seems that even if I set load_from and model/pre_trained the same checkpoint path, it errors out. Perhaps I am trying the wrong combinations?

Config File:-

%%writefile "/kaggle/working/Swin-Transformer-Object-Detection/configs/cascade_rcnn/cascade_mask_rcnn_swin_small_patch4_window7_mstrain_480-800_giou_4conv1f_adamw_3x_coco.py"
model = dict(
    type='CascadeRCNN',
    pretrained="/kaggle/working/checkpoints/swin_base_patch4_window7_224_22k.pth",
    backbone=dict(
        type='SwinTransformer',
        embed_dim=96,
        depths=[2, 2, 6, 2],          # [2, 2, 18, 2] --> [2,2,6,2] = 86M
        num_heads=[3, 6, 12, 24],         #reduce to [3,6,9,12]
        window_size=7,
        mlp_ratio=4.0,
        qkv_bias=True,
        qk_scale=None,
        drop_rate=0.0,
        attn_drop_rate=0.0,
        drop_path_rate=0.2,
        ape=False,
        patch_norm=True,
        out_indices=(0, 1, 2, 3),
        use_checkpoint=True),
    neck=dict(
        type='FPN',
        in_channels=[96, 192, 384, 768],
        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, 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=0.1111111111111111, 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='ConvFCBBoxHead',
                num_shared_convs=4,
                num_shared_fcs=1,
                in_channels=256,
                conv_out_channels=256,
                fc_out_channels=1024,
                roi_feat_size=7,
                num_classes=3,
                bbox_coder=dict(
                    type='DeltaXYWHBBoxCoder',
                    target_means=[0.0, 0.0, 0.0, 0.0],
                    target_stds=[0.1, 0.1, 0.2, 0.2]),
                reg_class_agnostic=False,
                reg_decoded_bbox=True,
                norm_cfg=dict(type='BN', requires_grad=True),
                loss_cls=dict(
                    type='CrossEntropyLoss',
                    use_sigmoid=False,
                    loss_weight=1.0),
                loss_bbox=dict(type='GIoULoss', loss_weight=10.0)),
            dict(
                type='ConvFCBBoxHead',
                num_shared_convs=4,
                num_shared_fcs=1,
                in_channels=256,
                conv_out_channels=256,
                fc_out_channels=1024,
                roi_feat_size=7,
                num_classes=3,
                bbox_coder=dict(
                    type='DeltaXYWHBBoxCoder',
                    target_means=[0.0, 0.0, 0.0, 0.0],
                    target_stds=[0.05, 0.05, 0.1, 0.1]),
                reg_class_agnostic=False,
                reg_decoded_bbox=True,
                norm_cfg=dict(type='BN', requires_grad=True),
                loss_cls=dict(
                    type='CrossEntropyLoss',
                    use_sigmoid=False,
                    loss_weight=1.0),
                loss_bbox=dict(type='GIoULoss', loss_weight=10.0)),
            dict(
                type='ConvFCBBoxHead',
                num_shared_convs=4,
                num_shared_fcs=1,
                in_channels=256,
                conv_out_channels=256,
                fc_out_channels=1024,
                roi_feat_size=7,
                num_classes=3,
                bbox_coder=dict(
                    type='DeltaXYWHBBoxCoder',
                    target_means=[0.0, 0.0, 0.0, 0.0],
                    target_stds=[0.033, 0.033, 0.067, 0.067]),
                reg_class_agnostic=False,
                reg_decoded_bbox=True,
                norm_cfg=dict(type='BN', requires_grad=True),
                loss_cls=dict(
                    type='CrossEntropyLoss',
                    use_sigmoid=False,
                    loss_weight=1.0),
                loss_bbox=dict(type='GIoULoss', loss_weight=10.0))
        ]),
    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_across_levels=False,
            nms_pre=2000,
            nms_post=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),
                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),
                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),
                pos_weight=-1,
                debug=False)
        ]),
    )
dataset_type = 'CocoDataset'
data_root = '/kaggle/working/mmdetection/data/coco'
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=False),
    dict(type='RandomFlip', flip_ratio=0.5),
    dict(
        type='AutoAugment',
        policies=[[{
            'type':
            'Resize',
            'img_scale': [(480, 1333), (512, 1333), (544, 1333), (576, 1333),
                          (608, 1333), (640, 1333), (672, 1333), (704, 1333),
                          (736, 1333), (768, 1333), (800, 1333)],
            'multiscale_mode':
            'value',
            'keep_ratio':
            True
        }],
                  [{
                      'type': 'Resize',
                      'img_scale': [(400, 1333), (500, 1333), (600, 1333)],
                      'multiscale_mode': 'value',
                      'keep_ratio': True
                  }, {
                      'type': 'RandomCrop',
                      'crop_type': 'absolute_range',
                      'crop_size': (384, 600),
                      'allow_negative_crop': True
                  }, {
                      'type':
                      'Resize',
                      'img_scale': [(480, 1333), (512, 1333), (544, 1333),
                                    (576, 1333), (608, 1333), (640, 1333),
                                    (672, 1333), (704, 1333), (736, 1333),
                                    (768, 1333), (800, 1333)],
                      'multiscale_mode':
                      'value',
                      'override':
                      True,
                      'keep_ratio':
                      True
                  }]]),
    dict(
        type='Normalize',
        mean=[123.675, 116.28, 103.53],
        std=[58.395, 57.12, 57.375],
        to_rgb=True),
    dict(type='Pad', size_divisor=32),
    dict(type='DefaultFormatBundle'),
    dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels'])
]
test_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        type='MultiScaleFlipAug',
        img_scale=(1333, 800),
        flip=False,
        transforms=[
            dict(type='Resize', keep_ratio=True),
            dict(type='RandomFlip'),
            dict(
                type='Normalize',
                mean=[123.675, 116.28, 103.53],
                std=[58.395, 57.12, 57.375],
                to_rgb=True),
            dict(type='Pad', size_divisor=32),
            dict(type='ImageToTensor', keys=['img']),
            dict(type='Collect', keys=['img'])
        ])
]
data = dict(
    samples_per_gpu=8,        # Batch_size ==> 4/8
    workers_per_gpu=4,
    train=dict(
        type='CocoDataset',
        classes=('fruit_woodiness', 'fruit_brownspot', 'fruit_healthy'),
        ann_file='/kaggle/working/mmdetection/data/coco/annotations/train.json',
        img_prefix='./ID',
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(type='LoadAnnotations', with_bbox=True, with_mask=False),
            dict(type='RandomFlip', flip_ratio=0.5),
            dict(
                type='AutoAugment',
                policies=[[{
                    'type':
                    'Resize',
                    'img_scale': [(480, 1333), (512, 1333), (544, 1333),
                                  (576, 1333), (608, 1333), (640, 1333),
                                  (672, 1333), (704, 1333), (736, 1333),
                                  (768, 1333), (800, 1333)],
                    'multiscale_mode':
                    'value',
                    'keep_ratio':
                    True
                }],
                          [{
                              'type': 'Resize',
                              'img_scale': [(400, 1333), (500, 1333),
                                            (600, 1333)],
                              'multiscale_mode': 'value',
                              'keep_ratio': True
                          }, {
                              'type': 'RandomCrop',
                              'crop_type': 'absolute_range',
                              'crop_size': (384, 600),
                              'allow_negative_crop': True
                          }, {
                              'type':
                              'Resize',
                              'img_scale': [(480, 1333), (512, 1333),
                                            (544, 1333), (576, 1333),
                                            (608, 1333), (640, 1333),
                                            (672, 1333), (704, 1333),
                                            (736, 1333), (768, 1333),
                                            (800, 1333)],
                              'multiscale_mode':
                              'value',
                              'override':
                              True,
                              'keep_ratio':
                              True
                          }]]),
            dict(
                type='Normalize',
                mean=[123.675, 116.28, 103.53],
                std=[58.395, 57.12, 57.375],
                to_rgb=True),
            dict(type='Pad', size_divisor=32),
            dict(type='DefaultFormatBundle'),
            dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels'])
        ],
        data_root='train'),
    val=dict(
        type='CocoDataset',
        classes=('fruit_woodiness', 'fruit_brownspot', 'fruit_healthy'),
        ann_file='/kaggle/working/mmdetection/data/coco/annotations/val.json',
        img_prefix='', #./ID
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                type='MultiScaleFlipAug',
                img_scale=(1333, 800),
                flip=False,
                transforms=[
                    dict(type='Resize', keep_ratio=True),
                    dict(type='RandomFlip'),
                    dict(
                        type='Normalize',
                        mean=[123.675, 116.28, 103.53],
                        std=[58.395, 57.12, 57.375],
                        to_rgb=True),
                    dict(type='Pad', size_divisor=32),
                    dict(type='ImageToTensor', keys=['img']),
                    dict(type='Collect', keys=['img'])
                ])
        ],
        data_root='val'),
    test=dict(
        type='CocoDataset',
        ann_file='/kaggle/working/mmdetection/data/coco/annotations/val.json',
        img_prefix='', #ID
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                type='MultiScaleFlipAug',
                img_scale=(1333, 800),
                flip=False,
                transforms=[
                    dict(type='Resize', keep_ratio=True),
                    dict(type='RandomFlip'),
                    dict(
                        type='Normalize',
                        mean=[123.675, 116.28, 103.53],
                        std=[58.395, 57.12, 57.375],
                        to_rgb=True),
                    dict(type='Pad', size_divisor=32),
                    dict(type='ImageToTensor', keys=['img']),
                    dict(type='Collect', keys=['img'])
                ])
        ]))
evaluation = dict(metric=['mAP', 'bbox'], classwise=True, interval=1)
optimizer = dict(
    type='AdamW',
    lr=0.0025,
    betas=(0.9, 0.999),
    weight_decay=0.05,
    paramwise_cfg=dict(
        custom_keys=dict(
            absolute_pos_embed=dict(decay_mult=0.0),
            relative_position_bias_table=dict(decay_mult=0.0),
            norm=dict(decay_mult=0.0))))
optimizer_config = dict(
    grad_clip=None,
    type='DistOptimizerHook',
    update_interval=1,
    coalesce=True,
    bucket_size_mb=-1,
    use_fp16=True)
lr_config = dict(
    policy='step',
    warmup='exponential',
    warmup_iters=500,
    warmup_ratio=0.333,
    step=[27, 33])
runner = dict(type='EpochBasedRunnerAmp', max_epochs=10)
checkpoint_config = dict(interval=2)
log_config = dict(interval=25, with_step=True, by_epoch=True, hooks=[dict(type='TextLoggerHook'), dict(type='TensorboardLoggerHook'), dict(type='WandbLoggerHook', init_kwargs=dict(project='makere-passion', name='neel'))])
custom_hooks = [dict(type='NumClassCheckHook')]
dist_params = dict(backend='nccl')
log_level = 'INFO'
load_from = "/kaggle/working/checkpoints/cascade_mask_rcnn_swin_small_patch4_window7.pth"
resume_from = None
workflow = [('train', 1)]
fp16 = None
work_dir = './results'
seed = 42
gpu_ids = 0
neel04 commented 2 years ago

Also, #250 is unsolved as well 

AttributeError: 'AmpState' object has no attribute 'loss_scalers'

Any ideas would be very, very much appreciated!