feat: add ctc segmentation tool to everyvoice #152
roedoejet
merged
21 commits into
main
from
dev.wav2vec2
feat: add ctc segmentation tool to everyvoice
e8283319
fix(reqs): bump to cuda 11.8
a035dbb8
fix: remove kaldi pitch detection
e402b830
fix: add sox to install for github actions
631e77bd
fix: only load audio as a string not path
416a9b5f
chore: update submodule
8deefd1b
fix: remove MMS model due to privacy concerns
5763ee2b
fix: add fix for hyphen and pipe in text
39980b8b
style(aligner): add pre-commit hooks
88bdefa1
fix: remove trailing and leading silence
2004a45e
fix: make 'no' the default option in wizard
2356b839
docs: add copy-able commands to train and preprocess help messages
d74c322a
refactor: remove output optional argument to preprocess step
de7d4f52
docs: add explanation to dataset and project name prompts
76f9a51a
fix: add training loss back to progress bar
c9758f84
fix: fix tests for wizard
bc45a59e
fix(fp): fix feature prediction scheduler
8763bcab
fix(docs): add index file again
a86a6b25
refactor: use set instead of list
00eb658c
fix: if cuda 11.8 is now required, make that the fresh env default
34d985e2
joanise
approved these changes
on 2023-11-17
docs: update readme to emphasize make-fresh-env, and other improvements
76e96bac
Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub