You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
hello.
is there limitation in model selection for training?
As I check in mmdet/models/detector/obb/ models with OAN are Faster_rcnn, orientedRCNN, roi_transformer and retinanet.
so should we select one of above for custome confile? and we cant use all models available in mmdetection?
this part I mean:
model = dict( type='oan_OrientedRCNN',
pretrained='torchvision://resnet50',
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'),
@mahilaMoghadami different models have their own attributes, you can read the realated papers and choose the best for your task. OAN could be added to mojarity plenty of existing models, like Oriented RCNN, Faster RCNN, RoI Trandformer and RetinaNet that we have uploaded, or you can add OAN to your custom models via referring to mmlab official documentation to change your models (https://mmdetection.readthedocs.io/en/latest/advanced_guides/customize_models.html).
hello.
is there limitation in model selection for training?
As I check in mmdet/models/detector/obb/ models with OAN are Faster_rcnn, orientedRCNN, roi_transformer and retinanet.
so should we select one of above for custome confile? and we cant use all models available in mmdetection?
this part I mean:
model = dict(
type='oan_OrientedRCNN',
pretrained='torchvision://resnet50',
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'),
oan = dict(
The text was updated successfully, but these errors were encountered: