feat: allow simple filenames when synthesizing - #154
Conversation
|
Review changes with Changed Files
|
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #154 +/- ##
=======================================
Coverage 48.80% 48.80%
=======================================
Files 27 27
Lines 2012 2012
Branches 231 231
=======================================
Hits 982 982
Misses 995 995
Partials 35 35 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
roedoejet
changed the title
|
Tests failed in CI on push. Can you fix that? |
| " (e.g. 'LJ050-0269.wav') instead of the default, which also includes" | ||
| " the speaker, language, and other metadata" | ||
| " (e.g. 'LJ050-0269--LJ--eng--ckpt=100000--v_ckpt=100000--pred.wav')." |
There was a problem hiding this comment.
Same comment as in StyleTTS2, I would remove the information that's redundant with what's shown in --output-dir. But yikes, I just reviewed the help for this command, the options are shown in an order that's not very logical. Out of scope for this PR, but we should do a thorough review of the order of all these synthesize parameters. text-to-wav has a better order, I would consult it in that review.
| vocoder_model=vocoder_model, | ||
| vocoder_config=vocoder_config, | ||
| vocoder_global_step=vocoder_global_step, | ||
| simple_filenames=simple_filenames, |
There was a problem hiding this comment.
This PR put in my face just how much repetition we have in this file. Out of scope for this PR, but it would be nice to find a way to factor how some of this stuff that just gets repeated over and over again.
8835153 to
d3a7809
Compare
2 participants
EveryVoiceTTS/EveryVoice#853