brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models module

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.C1(num_class=150, fc_dim=2048, use_softmax=False)

Bases: Module

forward(conv_out, segSize=None)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.C1DeepSup(num_class=150, fc_dim=2048, use_softmax=False)

Bases: Module

forward(conv_out, segSize=None)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.MobileNetV2Dilated(orig_net, dilate_scale=8)

Bases: Module

forward(x, return_feature_maps=False)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.ModelBuilder

Bases: object

static build_decoder(arch='ppm_deepsup', fc_dim=512, num_class=150, weights='', use_softmax=False)

static build_encoder(arch='resnet50dilated', fc_dim=512, weights='')

static weights_init(m)

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.PPM(num_class=150, fc_dim=4096, use_softmax=False, pool_scales=(1, 2, 3, 6))

Bases: Module

forward(conv_out, segSize=None)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.PPMDeepsup(num_class=150, fc_dim=4096, use_softmax=False, pool_scales=(1, 2, 3, 6))

Bases: Module

forward(conv_out, segSize=None)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.Resnet(orig_resnet)

Bases: Module

forward(x, return_feature_maps=False)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.ResnetDilated(orig_resnet, dilate_scale=8)

Bases: Module

forward(x, return_feature_maps=False)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.SegmentationModule(net_enc, net_dec, crit, deep_sup_scale=None)

Bases: SegmentationModuleBase

forward(feed_dict, *, segSize=None)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.SegmentationModuleBase

Bases: Module

pixel_acc(pred, label)

class brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.UPerNet(num_class=150, fc_dim=4096, use_softmax=False, pool_scales=(1, 2, 3, 6), fpn_inplanes=(256, 512, 1024, 2048), fpn_dim=256)

Bases: Module

forward(conv_out, segSize=None)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

brails.processors.FoundationClassifier.csail_segmentation_tool.csail_seg.models.models.conv3x3_bn_relu(in_planes, out_planes, stride=1)

3x3 convolution + BN + relu