Skip to content
This repository has been archived by the owner on Nov 22, 2022. It is now read-only.

Add custom metric class for reporting Joint model metrics #1339

Closed

Commits on Apr 29, 2020

  1. MultiLabel-MultiClass Model for Joint Sequence Tagging (facebookresea…

    …rch#1335)
    
    Summary:
    Pull Request resolved: facebookresearch#1335
    
    We need to support multi-class as well as multi-label prediction for joint models in pytext. This diff implements a
    
    1. Joint Multi Label Decoder
    2. MultiLabelClassification Output Layer
    3. Loss computation for multi-label-multi-class scenarios
    4. Label weights per label and per class
    5. Softmax options for output layers
    6. Custom Metric Reporter, Metric Class and Output for flow
    
    Differential Revision: D20210880
    
    fbshipit-source-id: 8d6292e0acd2ec8974bc8da4b86abc56f7b04f84
    shivanipods authored and facebook-github-bot committed Apr 29, 2020
    Copy the full SHA
    08339e2 View commit details
    Browse the repository at this point in the history
  2. Add custom metric class for reporting Joint model metrics (facebookre…

    …search#1339)
    
    Summary:
    Pull Request resolved: facebookresearch#1339
    
    Adding a multilabel metric class to support reporting all multi label and multi class metrics joint pytext models
    
    Reviewed By: seayoung1112
    
    Differential Revision: D21077306
    
    fbshipit-source-id: 3363f0faf55f18174c6e0c51e958681e6d3b4508
    shivanipods authored and facebook-github-bot committed Apr 29, 2020
    Copy the full SHA
    6c5118f View commit details
    Browse the repository at this point in the history