search

tryhiseyyysum / MENOL

This repo is about the official implementation of "MENOL: Multimodal-Enhanced Objectness Learner for Corner Case Detection in Autonomous Driving" (IEEE ICIP2024 oral paper)
Other
8 stars 0 forks source link

Replicating results with GOOD on CODA-VAL dataset #2

Open JoeWar99 opened 4 months ago

JoeWar99 commented 4 months ago

Hi there,

I would like to know what you did to achieve the results for the GOOD method on the CODA val dataset. I changed the model to be class aware and I tried to replicate the results but was only able to achieve an AR@100 of around 10%. Did you use additional augmentations beyond those already present in the GOOD repository config files? Did you change the score threshold or top_k hyperparemeters? I'm gonna leave my 2nd phase config files in this issue. Any help is appreciated.

Thank you.

# model settings
model = dict(
    type='FasterRCNN',
    backbone=dict(
        type='ResNet',
        depth=50,
        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',
        init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50')),
    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='L1Loss', loss_weight=1.0)),
    roi_head=dict(
        type='StandardRoIHead',
        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=7,
            bbox_coder=dict(
                type='DeltaXYWHBBoxCoder',
                target_means=[0., 0., 0., 0.],
                target_stds=[0.1, 0.1, 0.2, 0.2]),
            reg_class_agnostic=False,
            loss_cls=dict(
                type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
            loss_bbox=dict(type='L1Loss', 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=-1,
            pos_weight=-1,
            debug=False),
        rpn_proposal=dict(
            nms_pre=2000,
            max_per_img=1000,
            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)),
    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)
        # soft-nms is also supported for rcnn testing
        # e.g., nms=dict(type='soft_nms', iou_threshold=0.5, min_score=0.05)
    ))

dataset_type = 'SodaSplitDataset'
data_root_coda = 'CODA2022/'
data_root_soda = 'SODA10M/SSLAD-2D/labeled/'
img_norm_cfg = dict(
    mean=[103.743, 108.976, 110.09], std=[57.665, 61.197, 66.484], to_rgb=True)
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(
    type='AutoAugment',
    policies=[[
        dict(
            type='Resize',
            img_scale=[(480, 1280), (504, 1280), (528, 1280), (552, 1280),
                        (576, 1280), (600, 1280), (624, 1280), (648, 1280),
                        (672, 1280), (696, 1280), (720, 1280)],
            multiscale_mode='value',
            keep_ratio=True)
    ],
                [
                    dict(
                        type='Resize',
                        img_scale=[(400, 1280), (500, 1280), (600, 1280)],
                        multiscale_mode='value',
                        keep_ratio=True),
                    dict(
                        type='RandomCrop',
                        crop_type='absolute_range',
                        crop_size=(384, 600),
                        allow_negative_crop=True),
                    dict(
                        type='Resize',
                        img_scale=[(480, 1280), (504, 1280), (528, 1280), (552, 1280),
                                    (576, 1280), (600, 1280), (624, 1280), (648, 1280),
                                    (672, 1280), (696, 1280), (720, 1280)],
                        multiscale_mode='value',
                        override=True,
                        keep_ratio=True)
                ]]),
    dict(type='RandomFlip', flip_ratio=0.5),
    dict(type='Normalize', **img_norm_cfg),
    dict(type='Pad', size_divisor=1),
    dict(type='DefaultFormatBundle'),
    dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels'])
]

test_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        type='MultiScaleFlipAug',
        img_scale=(1280, 720),
        flip=False,
        transforms=[
            dict(type='Resize', keep_ratio=True),
            dict(type='RandomFlip'),
            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=2,
    workers_per_gpu=2,
    train=dict(
        type="SodaSplitPseudoBoxDataset",
        is_class_agnostic=False,
        train_class='voc',
        eval_class='nonvoc',
        pipeline=train_pipeline,
        ann_file=data_root_soda + 'annotations/instance_train_unknown.json',
        img_prefix=data_root_soda + 'train/',
        additional_ann_file=['good/work_dirs/pseudo_labels/soda_depth.json', 'good/work_dirs/pseudo_labels/soda_normal.json'],
        iou_thresh=0.5,
        score_thresh=0.7,
        top_k=1,
        random_sample_masks=False,
        merge_nms=True
        ),
    val=dict(
        is_class_agnostic=False,
        train_class='all',
        eval_class='all',
        ann_file=data_root_soda + 'annotations/instance_val_unknown.json',
        img_prefix=data_root_soda + 'val/',
        type=dataset_type,
        pipeline=test_pipeline),
    test=dict(
        is_class_agnostic=False,
        train_class='voc',
        eval_class='nonvoc',
        ann_file= data_root_coda + 'val/annotations_unknown.json',
        img_prefix= data_root_coda + "val/images/",
        type="CodaAwareDataset",
        pipeline=test_pipeline))

optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
optimizer_config = dict(grad_clip=None)
lr_config = dict(
    policy='step',
    warmup='linear',
    warmup_iters=500,
    warmup_ratio=0.001,
    step=[12, 14])
runner = dict(type='EpochBasedRunner', max_epochs=24)

checkpoint_config = dict(interval=2)
# yapf:disable
log_config = dict(
    interval=10,
    hooks=[
        dict(type='TextLoggerHook'),
        # dict(type='TensorboardLoggerHook')
    ])
# yapf:enable
dist_params = dict(backend='nccl')
log_level = 'INFO'
load_from = None
resume_from = None
workflow = [('train', 1)]
fsbarros98 commented 4 months ago

I'm having this issue too, can't replicate the paper results...