Skip to content
This repository has been archived by the owner on Oct 13, 2022. It is now read-only.

Use a pre-trained bigram P for LF-MMI training #222

Merged
merged 4 commits into from
Jul 2, 2021
Merged

Use a pre-trained bigram P for LF-MMI training #222

merged 4 commits into from
Jul 2, 2021

Commits on Jun 29, 2021

  1. Use a pre-trained bigram P for LF-MMI training

    @csukuangfj
    csukuangfj committed Jun 29, 2021
    Configuration menu
    Copy the full SHA
    c2c215b View commit details
    Browse the repository at this point in the history

Commits on Jun 30, 2021

  1. Fix errors for decoding.

    @csukuangfj
    csukuangfj committed Jun 30, 2021
    Configuration menu
    Copy the full SHA
    b80a3e8 View commit details
    Browse the repository at this point in the history

Commits on Jul 1, 2021

  1. Support loading models trained with P_scores.

    @csukuangfj
    csukuangfj committed Jul 1, 2021
    Configuration menu
    Copy the full SHA
    6632629 View commit details
    Browse the repository at this point in the history

  2. More fixes for checkpoints loading.

    @csukuangfj
    csukuangfj committed Jul 1, 2021
    Configuration menu
    Copy the full SHA
    42c3256 View commit details
    Browse the repository at this point in the history