here is how I have been getting through with custom backbone networks for resnet 101
class ModelResnet101FasterRCNN(FasterRCNN):
def __init__(self, data_conf, model_conf):
print("Creating model")
backbone_nn = torchvision.models.__dict__[model_conf["hyperParameters"]["net"]](pretrained=True)
# This line above yields the equivalent of ...
# backbone_nn = torchvision.models.resnet101(pretrained=True)
# OR
# backbone_nn = torchvision.models.wide_resnet101_2(pretrained=True)
modules = list(backbone_nn.children())[:-1] # delete the last fc layer.
backbone_nn = nn.Sequential(*modules)
for param in backbone_nn.parameters():
param.requires_grad = False
# FasterRCNN needs to know the number of
# output channels in a backbone. For resnet101, it's 2048
backbone_nn.out_channels = 2048
looks like you are not done yet… when using a custom backbone neural network the tutorial states you need to also instantiate the pooling components and attach them to your model… I am not seeing that in your notebook
from the tutorial you need…
# FasterRCNN needs to know the number of
# output channels in a backbone. For mobilenet_v2, it's 1280
# so we need to add it here
backbone.out_channels = 1280
# let's make the RPN generate 5 x 3 anchors per spatial
# location, with 5 different sizes and 3 different aspect
# ratios. We have a Tuple[Tuple[int]] because each feature
# map could potentially have different sizes and
# aspect ratios
anchor_generator = AnchorGenerator(sizes=((32, 64, 128, 256, 512),),
aspect_ratios=((0.5, 1.0, 2.0),))
# let's define what are the feature maps that we will
# use to perform the region of interest cropping, as well as
# the size of the crop after rescaling.
# if your backbone returns a Tensor, featmap_names is expected to
# be [0]. More generally, the backbone should return an
# OrderedDict[Tensor], and in featmap_names you can choose which
# feature maps to use.
roi_pooler = torchvision.ops.MultiScaleRoIAlign(featmap_names=[0],
output_size=7,
sampling_ratio=2)
# put the pieces together inside a FasterRCNN model
model = FasterRCNN(backbone,
num_classes=2,
rpn_anchor_generator=anchor_generator,
box_roi_pool=roi_pooler)
