licyk commited on
Commit
70c8d52
·
verified ·
1 Parent(s): f3dd414

Add the FCPE pitch model (fcpe/fcpe_c_v001.pt from torchfcpe 0.0.4)

Browse files

The model torchfcpe bundles, copied unchanged (SHA-256 b9aeaeb6…0015), with its MIT licence; manifest.json and README.md list it. rvc-next downloads it as the fcpe asset.

Files changed (4) hide show
  1. README.md +8 -3
  2. fcpe/LICENSE-FCPE.txt +21 -0
  3. fcpe/fcpe_c_v001.pt +3 -0
  4. manifest.json +10 -0
README.md CHANGED
@@ -14,9 +14,10 @@ RVC demo voices.
14
 
15
  Everything here is copied unchanged from the official RVC repository,
16
  [lj1995/VoiceConversionWebUI](https://huggingface.co/lj1995/VoiceConversionWebUI) (revision
17
- `e6d0c1a`). The demo voices come from inside its distribution packages (`RVC*.7z`), where they
18
- are the only copies. `manifest.json` lists every file with its size, SHA-256 and where it came
19
- from.
 
20
 
21
  ## Contents
22
 
@@ -24,6 +25,7 @@ from.
24
  | --- | --- | --- | --- |
25
  | `hubert/hubert_base/` | `config.json`, `preprocessor_config.json`, `pytorch_model.bin` | Content features (HuBERT/ContentVec), for conversion and training | `hubert_base/` |
26
  | `rmvpe/` | `rmvpe.pt`, `rmvpe.onnx` | Pitch extraction (RMVPE); the ONNX export for DirectML | `rmvpe.pt`, `rmvpe.onnx` |
 
27
  | `pretrained/v1/` | `{f0,}{G,D}{32k,40k,48k}.pth` | Base models for training v1 voices, with (`f0`) and without pitch guidance | `pretrained/` |
28
  | `pretrained/v2/` | the same twelve files | Base models for training v2 voices | `pretrained_v2/` |
29
  | `separation/` | five BS-/Mel-Roformer checkpoints and their YAML | Vocal separation, dereverb and karaoke (PyMSS) | `pymss_weights/` |
@@ -50,3 +52,6 @@ software is for research use only, and users bear full responsibility for the vo
50
  and distribute; anyone who does not accept this may not use these files. The licences of the
51
  components the models build on (ContentVec, VITS, HiFi-GAN, UVR, audio-slicer) are listed in the
52
  same file.
 
 
 
 
14
 
15
  Everything here is copied unchanged from the official RVC repository,
16
  [lj1995/VoiceConversionWebUI](https://huggingface.co/lj1995/VoiceConversionWebUI) (revision
17
+ `e6d0c1a`), except the FCPE pitch model, copied unchanged from the
18
+ [torchfcpe](https://github.com/CNChTu/FCPE) 0.0.4 package, which bundles it. The demo voices come
19
+ from inside RVC's distribution packages (`RVC*.7z`), where they are the only copies.
20
+ `manifest.json` lists every file with its size, SHA-256 and where it came from.
21
 
22
  ## Contents
23
 
 
25
  | --- | --- | --- | --- |
26
  | `hubert/hubert_base/` | `config.json`, `preprocessor_config.json`, `pytorch_model.bin` | Content features (HuBERT/ContentVec), for conversion and training | `hubert_base/` |
27
  | `rmvpe/` | `rmvpe.pt`, `rmvpe.onnx` | Pitch extraction (RMVPE); the ONNX export for DirectML | `rmvpe.pt`, `rmvpe.onnx` |
28
+ | `fcpe/` | `fcpe_c_v001.pt`, `LICENSE-FCPE.txt` | Pitch extraction (FCPE) | `torchfcpe/assets/fcpe_c_v001.pt` in torchfcpe 0.0.4 |
29
  | `pretrained/v1/` | `{f0,}{G,D}{32k,40k,48k}.pth` | Base models for training v1 voices, with (`f0`) and without pitch guidance | `pretrained/` |
30
  | `pretrained/v2/` | the same twelve files | Base models for training v2 voices | `pretrained_v2/` |
31
  | `separation/` | five BS-/Mel-Roformer checkpoints and their YAML | Vocal separation, dereverb and karaoke (PyMSS) | `pymss_weights/` |
 
52
  and distribute; anyone who does not accept this may not use these files. The licences of the
53
  components the models build on (ContentVec, VITS, HiFi-GAN, UVR, audio-slicer) are listed in the
54
  same file.
55
+
56
+ The FCPE model (`fcpe/`) is not part of RVC: it is MIT-licensed by its authors (Copyright (c) 2023
57
+ CN_ChiTu), without the research-use clause; its licence is `fcpe/LICENSE-FCPE.txt`.
fcpe/LICENSE-FCPE.txt ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MIT License
2
+
3
+ Copyright (c) 2023 CN_ChiTu
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
fcpe/fcpe_c_v001.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9aeaeb673436eeda50ceafd632aa681aa63417e52eae4207503d180c9b10015
3
+ size 43363260
manifest.json CHANGED
@@ -294,6 +294,16 @@
294
  "path": "rmvpe.pt"
295
  }
296
  },
 
 
 
 
 
 
 
 
 
 
297
  {
298
  "path": "separation/config_mel_band_roformer_karaoke.yaml",
299
  "category": "separation",
 
294
  "path": "rmvpe.pt"
295
  }
296
  },
297
+ {
298
+ "path": "fcpe/fcpe_c_v001.pt",
299
+ "category": "fcpe",
300
+ "size": 43363260,
301
+ "sha256": "b9aeaeb673436eeda50ceafd632aa681aa63417e52eae4207503d180c9b10015",
302
+ "origin": {
303
+ "repository": "pypi:torchfcpe==0.0.4",
304
+ "path": "torchfcpe/assets/fcpe_c_v001.pt"
305
+ }
306
+ },
307
  {
308
  "path": "separation/config_mel_band_roformer_karaoke.yaml",
309
  "category": "separation",