-
feat: add ctc segmentation tool to everyvoice
-
fix(reqs): bump to cuda 11.8
-
fix: remove kaldi pitch detection
-
fix: add sox to install for github actions
-
fix: only load audio as a string not path
-
chore: update submodule
-
fix: remove MMS model due to privacy concerns
-
fix: add fix for hyphen and pipe in text
-
style(aligner): add pre-commit hooks
-
fix: remove trailing and leading silence
-
fix: make 'no' the default option in wizard
-
docs: add copy-able commands to train and preprocess help messages
-
refactor: remove output optional argument to preprocess step
-
docs: add explanation to dataset and project name prompts
-
fix: add training loss back to progress bar
-
fix: fix tests for wizard
-
fix(fp): fix feature prediction scheduler
-
fix(docs): add index file again
-
refactor: use set instead of list
-
fix: if cuda 11.8 is now required, make that the fresh env default
joanise committed on 2023-11-17 21:03
-
docs: update readme to emphasize make-fresh-env, and other improvements