Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Major Refactor #14

Merged
merged 8 commits into from
Oct 14, 2023
Merged

Major Refactor #14

merged 8 commits into from
Oct 14, 2023

Commits on Oct 3, 2023

  1. Make datasets preprocessing scalable for large datasets.

    @harrykeightley
    harrykeightley committed Oct 3, 2023
    Configuration menu
    Copy the full SHA
    12b6696 View commit details
    Browse the repository at this point in the history

Commits on Oct 5, 2023

  1. Improve large dataset performance.

    @harrykeightley
    harrykeightley committed Oct 5, 2023
    Configuration menu
    Copy the full SHA
    b4d438b View commit details
    Browse the repository at this point in the history

Commits on Oct 9, 2023

  1. Make model-specific arguments flexible, and specifiable in the training 

    job.
- Explicitly set up optimizer.
    @harrykeightley
    harrykeightley committed Oct 9, 2023
    Configuration menu
    Copy the full SHA
    06a11d3 View commit details
    Browse the repository at this point in the history

  2. Fix bug where replacing vocab character with the same character would 

    cause it to remove that character entirely from the vocab.
    @harrykeightley
    harrykeightley committed Oct 9, 2023
    Configuration menu
    Copy the full SHA
    7a64f0f View commit details
    Browse the repository at this point in the history

  3. Cleanup dataset processing 

    - Force absolute audio path resolution for dataset files
- Manually load in audio from paths with librosa
- More logging

.
    @harrykeightley
    harrykeightley committed Oct 9, 2023
    Configuration menu
    Copy the full SHA
    0d0c864 View commit details
    Browse the repository at this point in the history

  4. Bad attempt at upgrading all dependencies

    @harrykeightley
    harrykeightley committed Oct 9, 2023
    Configuration menu
    Copy the full SHA
    15ed847 View commit details
    Browse the repository at this point in the history

Commits on Oct 14, 2023

  1. Trainer upgrades (#13) 

    * Refactor trainer code to expose as much flexibility to user as possible w.r.t. training options. 
* Update tests
* Finally get a working wav2vec2 training run.
* Add example script for working wav2vec2 run.
    @harrykeightley
    harrykeightley authored Oct 14, 2023
    Configuration menu
    Copy the full SHA
    f3e3a07 View commit details
    Browse the repository at this point in the history

  2. Update version and format.

    @harrykeightley
    harrykeightley committed Oct 14, 2023
    Configuration menu
    Copy the full SHA
    c8b40b9 View commit details
    Browse the repository at this point in the history