Update README.md
Browse files
README.md
CHANGED
@@ -10,31 +10,15 @@ metrics:
|
|
10 |
![model-cover.png](https://huggingface.co/ArkanDash/rvc-genshin-impact/resolve/main/model-cover.png)
|
11 |
|
12 |
# About Retrieval based Voice Conversion (RVC)
|
13 |
-
Learn more about Retrieval based Voice Conversion in this link below:<br />
|
14 |
-
[RVC WebUI](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI)
|
15 |
-
|
16 |
-
# How to use?
|
17 |
-
Download the prezipped model and put to your RVC Project.
|
18 |
-
|
19 |
-
Model test: [Google Colab](https://colab.research.google.com/drive/110kiMZTdP6Ri1lY9-NbQf17GVPPhHyeT?usp=sharing) / [RVC Models New](https://huggingface.co/spaces/ArkanDash/rvc-models-new) (Which is basically the same but hosted on spaces)
|
20 |
|
21 |
|
22 |
## <center> INFO <br />
|
23 |
Model Created by ArkanDash <br />
|
24 |
-
The voice that was used in this model belongs to Hoyoverse.
|
25 |
|
26 |
-
The voice I make to make this model was ripped from the game (3.7). <br />
|
27 |
-
Total Models: 28 Models (19 V1 Models & 9 V2 Models)
|
28 |
|
29 |
### V1 Model <br />
|
30 |
This was trained on Original RVC.<br />
|
31 |
-
Pitch Extract using Harvest.<br />
|
32 |
-
This model was trained with 100 epochs, 10 batch sizes, and a 40K sample rate (some models had a 48k sample rate).<br />
|
33 |
-
Every V1 model was trained more or less around 30 minutes of character voice.
|
34 |
|
35 |
-
I may exclude some models to higher epochs due to the low duration of the character's voice.<br />
|
36 |
-
- Klee 150 Epochs
|
37 |
-
- Fischl 150 Epochs
|
38 |
|
39 |
### (New) V2 Model <br />
|
40 |
This was trained on Mangio-Fork RVC.<br />
|
@@ -42,11 +26,3 @@ Pitch Extract using Crepe.<br />
|
|
42 |
This model was trained with 100 epochs, 8 batch sizes, and a 48K sample rate. (some models had a 40k sample rate).<br />
|
43 |
Every V2 model was trained more or less around 60 minutes of character voice.
|
44 |
|
45 |
-
Other request:<br />
|
46 |
-
- Greater Lord Rukkhadevata: 750 Epochs, 16 Batch size, 48k Sample rate. (10 minutes dataset)
|
47 |
-
- Charlotte: 400 Epochs, 16 Batch size, 48k Sample rate. (18 minutes dataset)
|
48 |
-
|
49 |
-
Note:
|
50 |
-
- For faruzan, somehow the index file is smaller, But it output a log when training here: <br />
|
51 |
-
`Converged (lack of improvement in inertia) at step 1152/48215` <br />
|
52 |
-
Let me know if you have an issue.
|
|
|
10 |
![model-cover.png](https://huggingface.co/ArkanDash/rvc-genshin-impact/resolve/main/model-cover.png)
|
11 |
|
12 |
# About Retrieval based Voice Conversion (RVC)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
|
14 |
|
15 |
## <center> INFO <br />
|
16 |
Model Created by ArkanDash <br />
|
|
|
17 |
|
|
|
|
|
18 |
|
19 |
### V1 Model <br />
|
20 |
This was trained on Original RVC.<br />
|
|
|
|
|
|
|
21 |
|
|
|
|
|
|
|
22 |
|
23 |
### (New) V2 Model <br />
|
24 |
This was trained on Mangio-Fork RVC.<br />
|
|
|
26 |
This model was trained with 100 epochs, 8 batch sizes, and a 48K sample rate. (some models had a 40k sample rate).<br />
|
27 |
Every V2 model was trained more or less around 60 minutes of character voice.
|
28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|