Add Faster RCNN Configuration using EfficientNet backbone

[TNodeCode]

Describe the feature

The following configuration file uses an EfficientNet backbone from the mmpretrain library (which must be installed in the Python environment to make this work).

Motivation
In my experiments EfficientNet backbones often achieve higher mAP50 scores on my datasets than using ResNet backbones.

The configuration

location: configs/faster_rcnn/faster-rcnn_enxl_fpn_1x_coco.py

_base_ = [
    '../_base_/models/faster-rcnn_r50_fpn.py',
    '../_base_/datasets/coco_detection.py',
    '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
]

_base_.model.backbone=dict(
    type='mmpretrain.EfficientNetV2', # Using EfficientNetV2 from mmpretrain
    arch='xl',
    frozen_stages=1,
    out_indices=(2,3,5,8), 
    init_cfg=dict(
        type='Pretrained',
        checkpoint='https://download.openmmlab.com/mmclassification/v0/efficientnetv2/efficientnetv2-xl_in21k-pre-3rdparty_in1k_20221220-583ac18b.pth',
        prefix='backbone.')
)
_base_.model.neck=dict(
    type='FPN',
    in_channels=[64, 96, 256, 1280],
    out_channels=256,
    num_outs=5
)

[lauriebax]

Cool. Would you be able to create a PR with the new config?