deepdisc.model.models

Classes

WeightedRedshiftPDFCasROIHeads	CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for
CNNRedshiftPDFCasROIHeads	CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for
RedshiftPDFCasROIHeadsGold	CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for
RedshiftPDFCasROIHeads	CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for
OldRedshiftPDFCasROIHeads	CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for
RedshiftPointCasROIHeads	CascadeROIHeads with added redshift point estimate capability. Follows the detectron2 CascadeROIHeads class init
RedshiftPointROIHeads	ROIHead with added redshift point estimate capability. Follows the detectron2 StandardROIHead class init
RedshiftPDFROIHeads	ROIHead with added redshift pdf capability. Follows the detectron2 StandardROIHead class init, except for
OldEBVRedshiftPDFCasROIHeads	CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for

Functions

return_lazy_model(cfg[, freeze])

Return a model formed from a LazyConfig with the backbone

Module Contents

return_lazy_model(cfg, freeze=True)

Return a model formed from a LazyConfig with the backbone frozen. Only the head layers will be trained.

Parameters:

cfg (.py file) – a LazyConfig

Return type:

torch model

class WeightedRedshiftPDFCasROIHeads(num_components: int, zloss_factor: float, weights: List[float], zbins: List[float], *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_heads: List[torch.nn.Module], box_predictors: List[torch.nn.Module], proposal_matchers: List[detectron2.modeling.matcher.Matcher], **kwargs)

Bases: detectron2.modeling.roi_heads.CascadeROIHeads

CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for

Parameters:

num_components (int) – Number of gaussian components in the Mixture Density Network

redshift_pooler

_output_size

num_components

zloss_factor

redshift_fc

output_pdf(inputs)

_forward_redshift(features, instances)

forward(images, features, proposals, targets=None)

class CNNRedshiftPDFCasROIHeads(num_components: int, zloss_factor: float, *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_heads: List[torch.nn.Module], box_predictors: List[torch.nn.Module], proposal_matchers: List[detectron2.modeling.matcher.Matcher], **kwargs)

Bases: detectron2.modeling.roi_heads.CascadeROIHeads

CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for

Parameters:

num_components (int) – Number of gaussian components in the Mixture Density Network

redshift_pooler

_output_size

num_components

zloss_factor

redshift_conv

output_pdf(inputs)

_forward_redshift(features, instances)

forward(images, features, proposals, targets=None)

class RedshiftPDFCasROIHeadsGold(num_components: int, zloss_factor: float, *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_heads: List[torch.nn.Module], box_predictors: List[torch.nn.Module], proposal_matchers: List[detectron2.modeling.matcher.Matcher], **kwargs)

Bases: detectron2.modeling.roi_heads.CascadeROIHeads

CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for

Parameters:

num_components (int) – Number of gaussian components in the Mixture Density Network

redshift_pooler

_output_size

num_components

zloss_factor

redshift_fc

output_pdf(inputs)

_forward_redshift(features, instances)

forward(images, features, proposals, targets=None)

class RedshiftPDFCasROIHeads(num_components: int, zloss_factor: float, *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_heads: List[torch.nn.Module], box_predictors: List[torch.nn.Module], proposal_matchers: List[detectron2.modeling.matcher.Matcher], **kwargs)

Bases: detectron2.modeling.roi_heads.CascadeROIHeads

CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for

Parameters:

num_components (int) – Number of gaussian components in the Mixture Density Network

redshift_pooler

_output_size

num_components

zloss_factor

redshift_fc

output_pdf(inputs)

_forward_redshift(features, instances)

forward(images, features, proposals, targets=None)

class OldRedshiftPDFCasROIHeads(num_components: int, zloss_factor: float, *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_heads: List[torch.nn.Module], box_predictors: List[torch.nn.Module], proposal_matchers: List[detectron2.modeling.matcher.Matcher], **kwargs)

Bases: detectron2.modeling.roi_heads.CascadeROIHeads

CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for

Parameters:

num_components (int) – Number of gaussian components in the Mixture Density Network

redshift_pooler

_output_size

num_components

zloss_factor

redshift_fc

output_pdf(inputs)

_forward_redshift(features, instances)

forward(images, features, proposals, targets=None)

class RedshiftPointCasROIHeads(zloss_factor: float, *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_heads: List[torch.nn.Module], box_predictors: List[torch.nn.Module], proposal_matchers: List[detectron2.modeling.matcher.Matcher], **kwargs)

Bases: detectron2.modeling.roi_heads.CascadeROIHeads

CascadeROIHeads with added redshift point estimate capability. Follows the detectron2 CascadeROIHeads class init

redshift_pooler

_output_size

zloss_factor

redshift_fc

_forward_redshift(features, instances)

forward(images, features, proposals, targets=None)

class RedshiftPointROIHeads(zloss_factor: float, *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_head: torch.nn.Module, box_predictor: torch.nn.Module, mask_in_features: List[str] | None = None, mask_pooler: detectron2.modeling.poolers.ROIPooler | None = None, mask_head: torch.nn.Module | None = None, keypoint_in_features: List[str] | None = None, keypoint_pooler: detectron2.modeling.poolers.ROIPooler | None = None, keypoint_head: torch.nn.Module | None = None, train_on_pred_boxes: bool = False, **kwargs)

Bases: detectron2.modeling.roi_heads.StandardROIHeads

ROIHead with added redshift point estimate capability. Follows the detectron2 StandardROIHead class init

redshift_pooler

_output_size

redshift_fc

_forward_redshift(features, instances)

forward(images: detectron2.structures.ImageList, features: Dict[str, torch.Tensor], proposals: List[detectron2.structures.Instances], targets: List[detectron2.structures.Instances] | None = None) → Tuple[List[detectron2.structures.Instances], Dict[str, torch.Tensor]]

See ROIHeads.forward.

class RedshiftPDFROIHeads(num_components: int, zloss_factor: float, *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_head: torch.nn.Module, box_predictor: torch.nn.Module, mask_in_features: List[str] | None = None, mask_pooler: detectron2.modeling.poolers.ROIPooler | None = None, mask_head: torch.nn.Module | None = None, keypoint_in_features: List[str] | None = None, keypoint_pooler: detectron2.modeling.poolers.ROIPooler | None = None, keypoint_head: torch.nn.Module | None = None, train_on_pred_boxes: bool = False, **kwargs)

Bases: detectron2.modeling.roi_heads.StandardROIHeads

ROIHead with added redshift pdf capability. Follows the detectron2 StandardROIHead class init, except for

Parameters:

num_components (int) – Number of gaussian components in the Mixture Density Network

redshift_pooler

zloss_factor

_output_size

num_components

redshift_fc

output_pdf(inputs)

_forward_redshift(features, instances)

forward(images: detectron2.structures.ImageList, features: Dict[str, torch.Tensor], proposals: List[detectron2.structures.Instances], targets: List[detectron2.structures.Instances] | None = None) → Tuple[List[detectron2.structures.Instances], Dict[str, torch.Tensor]]

See ROIHeads.forward.

class OldEBVRedshiftPDFCasROIHeads(num_components: int, zloss_factor: float, *, box_in_features: List[str], box_pooler: detectron2.modeling.poolers.ROIPooler, box_heads: List[torch.nn.Module], box_predictors: List[torch.nn.Module], proposal_matchers: List[detectron2.modeling.matcher.Matcher], **kwargs)

Bases: detectron2.modeling.roi_heads.CascadeROIHeads

CascadeROIHead with added redshift pdf capability. Follows the detectron2 CascadeROIHead class init, except for

Parameters:

num_components (int) – Number of gaussian components in the Mixture Density Network

redshift_pooler

_output_size

num_components

zloss_factor

redshift_fc

output_pdf(inputs)

_forward_redshift(features, instances)

forward(images, features, proposals, targets=None)