[Fix] Albumentations default key mapping mismatch

[Alias-z]

Modification

Fix Albumentations default key mapping mismatch as mentioned in issue # 3179 by changing self.keymap_to_albu = { 'img': 'image', 'gt_masks': 'masks'} to self.keymap_to_albu = { 'img': 'image', 'gt_seg_map': 'mask'}

Use cases (Optional)

Example albu config

crop_size = (512, 512)

albu_train_transforms = [
    dict(
        type='PadIfNeeded',
        min_height=crop_size[0]*2,
        min_width=crop_size[1]*2,
        border_mode=0,
        always_apply=True),
    dict(type='Flip', always_apply=True),
    dict(type='Rotate', limit=(-180, 180), interpolation=4, always_apply=True),
    dict(type='RandomScale', scale_limit=0.1, interpolation=4, always_apply=True),
    dict(
        type='ElasticTransform',
        alpha=20, 
        sigma=15, 
        interpolation=4, 
        border_mode=0,
        mask_value=(0, 0, 0),
        approximate=True,
        same_dxdy=True,
        p=0.8),
    dict(type='ColorJitter', brightness=0.2, contrast=0.1, saturation=0.2, hue=0.2, always_apply=True),
    dict(type='AdvancedBlur', p=0.5),
    dict(type='CenterCrop', height=crop_size[0], width=crop_size[1], always_apply=True)
]

Example training pipeline without specifying keymap

train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations', reduce_zero_label=False),
    dict(
        type='Albu',
        transforms=albu_train_transforms,
        ),
    dict(type='Resize', scale=crop_size, keep_ratio=False, interpolation='lanczos'), 
    dict(type='PackSegInputs')
]

Example viz_dataset before the issue fixing

Example viz_dataset after the issue fixing

[CLAassistant]

All committers have signed the CLA.

[OpenMMLab-Assistant-004]

Hi @Alias-z,

We'd like to express our appreciation for your valuable contributions to the mmsegmentation. Your efforts have significantly aided in enhancing the project's quality.
It is our pleasure to invite you to join our community thorugh Discord_Special Interest Group (SIG) channel. This is a great place to share your experiences, discuss ideas, and connect with other like-minded people. To become a part of the SIG channel, send a message to the moderator, OpenMMLab, briefly introduce yourself and mention your open-source contributions in the #introductions channel. Our team will gladly facilitate your entry. We eagerly await your presence. Please follow this link to join us: ​https://discord.gg/UjgXkPWNqA.

If you're on WeChat, we'd also love for you to join our community there. Just add our assistant using the WeChat ID: openmmlabwx. When sending the friend request, remember to include the remark "mmsig + Github ID".

Thanks again for your awesome contribution, and we're excited to have you as part of our community!