Skip to content

Question about Confident Learning #1

Description

@icecity96

According to your paper, sample weight should be Var(y)

In sampleer.py line 76:

for label in range(num_classes):
      # .....
       w_s.extend(y_var_[indices][:,0])

Why there always take class0's confidence?

Does this line should be:

w_s.extend(y_var_[indices][:, label]

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions