add initial README and models
Browse files- README.md +47 -3
- brickwall-e32.ckpt +3 -0
- brickwall_e96_a168096.pth +3 -0
README.md
CHANGED
@@ -1,3 +1,47 @@
|
|
1 |
-
---
|
2 |
-
license: mit
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
language:
|
4 |
+
- ja
|
5 |
+
pipeline_tag: text-to-speech
|
6 |
+
---
|
7 |
+
Following this guide with exceptions
|
8 |
+
https://rentry.org/GPT-SoVITS-guide
|
9 |
+
|
10 |
+
I used the latest git pull from
|
11 |
+
https://github.com/RVC-Boss/GPT-SoVITS/
|
12 |
+
|
13 |
+
I needed to put:
|
14 |
+
|
15 |
+
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/path/to/cudann/lib/
|
16 |
+
|
17 |
+
in my shell.
|
18 |
+
|
19 |
+
|
20 |
+
put the pth file in your SoVITS_weights_v2 folder and the ckpt in GPT_weights_v2
|
21 |
+
|
22 |
+
make both Language for Reference audio and Inference text language "Japanese", set slicing to "Slice by every punct".
|
23 |
+
|
24 |
+
the underlying model was trained on audio that was brickwall dynamic range compressed, like that commonly found in visual novels or video games.
|
25 |
+
|
26 |
+
you should be able to give it a CLEAN AND NOISE/MUSIC/STATIC free FEMALE Japanese voice clip from 3-10 seconds, give it a 100% ACCURATE transcription and get ok results out the other side.
|
27 |
+
|
28 |
+
I have found that results can be improved applying post-generation noise reduction and some treble boosting EQ processing. Audacity works well enough for this, since there isn't anything in the official Gradio interface.
|
29 |
+
|
30 |
+
Feel free to keep everything else at the deaults
|
31 |
+
|
32 |
+
If you want to start the inference engine auomatically, you can use do something like
|
33 |
+
|
34 |
+
python3 /path/to/GPT_SoVITS/inference_webui.py "Auto"
|
35 |
+
|
36 |
+
If you isolate it ala https://rentry.org/IsolatedLinuxWebService and put nginx in front of it with an ssl cert, you need something like this in the location block:
|
37 |
+
|
38 |
+
proxy_pass http://127.0.0.1:9872/;
|
39 |
+
proxy_buffering off;
|
40 |
+
proxy_redirect off;
|
41 |
+
proxy_http_version 1.1;
|
42 |
+
proxy_set_header Upgrade $http_upgrade;
|
43 |
+
proxy_set_header Connection "upgrade";
|
44 |
+
proxy_set_header Host $host;
|
45 |
+
client_max_body_size 500M;
|
46 |
+
proxy_set_header X-Forwarded-Proto $scheme;
|
47 |
+
add_header 'Content-Security-Policy' 'upgrade-insecure-requests';
|
brickwall-e32.ckpt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ab384212e78adbefb59aca1cc99acc910f896eb43e5c3b6a031ccf5f67576631
|
3 |
+
size 155312957
|
brickwall_e96_a168096.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:19bb170ff1e2cf75c7be5235ee9b68210f7f48dd9d627267f96abb3068d89950
|
3 |
+
size 85007202
|