Merge branch 'main' of https://huggingface.co/spaces/dthomas84/RVC_RULE1
Browse files
app.py
CHANGED
@@ -1573,13 +1573,17 @@ with gr.Blocks(title='RVC First Rule v1', theme=theme) as app:
|
|
1573 |
be vocals only, preferable under 30 seconds in length. Longer clips
|
1574 |
can be cloned, but will take longer. Second, a model needs to be chosen --
|
1575 |
either Jenny's ("jenny.pth") or Joel's ("joel.pth"). If you do not see them as options,
|
1576 |
-
make sure to "Refresh" the model choices.
|
1577 |
-
|
|
|
1578 |
This is especially useful for male to female or female to male conversions
|
1579 |
-
adjustment settings.
|
1580 |
-
to male conversions. There are other more advanced settings, including proportion
|
1581 |
of features retrieved, which can be used to adjust how closely the algorithm with
|
1582 |
-
match features from one voice to another.
|
|
|
|
|
|
|
|
|
1583 |
but you can certainly play around with this to get different results. Most of all,
|
1584 |
have fun! </h3>
|
1585 |
""")
|
|
|
1573 |
be vocals only, preferable under 30 seconds in length. Longer clips
|
1574 |
can be cloned, but will take longer. Second, a model needs to be chosen --
|
1575 |
either Jenny's ("jenny.pth") or Joel's ("joel.pth"). If you do not see them as options,
|
1576 |
+
make sure to "Refresh" the model choices.
|
1577 |
+
|
1578 |
+
Initially, simply clone the clip to set a baseline. Then, you can play around with the adjustment settings.
|
1579 |
This is especially useful for male to female or female to male conversions
|
1580 |
+
adjustment settings. There are other more advanced settings, including proportion
|
|
|
1581 |
of features retrieved, which can be used to adjust how closely the algorithm with
|
1582 |
+
match features from one voice to another.
|
1583 |
+
|
1584 |
+
If you receive an error message upon uploading an audio file or cloning a voice,
|
1585 |
+
hit "Refresh" and then try again. As this is as an early iteration and bleeding edge tech,
|
1586 |
+
there can be bugs. The default settings usually work well,
|
1587 |
but you can certainly play around with this to get different results. Most of all,
|
1588 |
have fun! </h3>
|
1589 |
""")
|