[Feature] Add Base Retriever and Image2Image Retriever for retrieval tasks

[zzc98]

Motivation

• A base class created to support generic retrieval tasks.
• Realize the image to image retriever.

Use cases

gallery_dataloader = dict(
    batch_size=8,
    dataset=dict(
        type='InShop',
        data_root='data/inshop',
        ann_file='list_eval_partition.txt',
        mode='gallery',
        pipeline=val_pipeline
    ),
    sampler=dict(type='DefaultSampler', shuffle=False))

model = dict(
    type='ImageToImageRetriever',
    image_encoder=[
        dict(type='ResNet',
                  depth=101,
                  num_stages=4,
                  out_indices=(3,),
                  style='pytorch'),
        dict(type='GeneralizedMeanPooling'),
        dict(type='LinearReduction',
               in_channels=2048,
               out_channels=512,
               norm_cfg=dict(type='BN1d'),
               act_cfg=dict(type='ReLU'))
    ],
    head=dict(type='ArcFaceHead',
              num_classes=3997,
              in_channels=512,
              init_cfg=dict(type='Xavier', layer='Linear', distribution='uniform'),
              loss=dict(type='CrossEntropyLoss', loss_weight=1.0)),
    prototype=gallery_dataloader
)

Note:

• The implement of InShop can be found in [Feature] Support InShop Dataset (Image Retrieval). #1019.
• The implement of LinearReduction can be found in [Feature] add reduction for neck #978.
• The implement of ArcFaceHead can be found in Add arcface head #926.

Checklist

Before PR:

• Pre-commit or other linting tools are used to fix the potential lint issues.
• Bug fixes are fully covered by unit tests, the case that causes the bug should be added in the unit tests.
• The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
• The documentation has been modified accordingly, like docstring or example tutorials.

After PR:

• If the modification has potential influence on downstream or other related projects, this PR should be tested with those projects, like MMDet or MMSeg.
• CLA has been signed and all committers have signed the CLA in this PR.

[Ezra-Yu]

please fix the unit tests.

[codecov]

Codecov Report

Base: 0.02% // Head: 91.51% // Increases project coverage by +91.48% 🎉

Coverage data is based on head (efe2f1a) compared to base (b8b31e9).
Patch has no changes to coverable lines.

❗ Current head efe2f1a differs from pull request most recent head 937122a. Consider uploading reports for the commit 937122a to get more accurate results

Additional details and impacted files

@@             Coverage Diff              @@
##           dev-1.x    #1055       +/-   ##
============================================
+ Coverage     0.02%   91.51%   +91.48%     
============================================
  Files          121      135       +14     
  Lines         8217    10120     +1903     
  Branches      1368     1630      +262     
============================================
+ Hits             2     9261     +9259     
+ Misses        8215      654     -7561     
- Partials         0      205      +205     

Flag	Coverage Δ
unittests	91.51% <ø> (+91.48%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
mmcls/apis/inference.py	0.00% <0.00%> (ø)
mmcls/datasets/transforms/compose.py
mmcls/models/backbones/efficientformer.py	95.08% <0.00%> (ø)
mmcls/structures/utils.py	77.77% <0.00%> (ø)
mmcls/models/retrievers/base.py	100.00% <0.00%> (ø)
mmcls/engine/hooks/retriever_hooks.py	72.72% <0.00%> (ø)
mmcls/models/backbones/mobilevit.py	91.15% <0.00%> (ø)
mmcls/models/retrievers/image2image.py	92.38% <0.00%> (ø)
mmcls/models/backbones/edgenext.py	95.20% <0.00%> (ø)
mmcls/models/backbones/mvit.py	92.46% <0.00%> (ø)
... and 125 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[Ezra-Yu]

LGTM.