Comments (4)
(knn) root@fangg-MS-7B78:/home/fangg/tts/knn-vc-master# python prematch_dataset.py --librispeech_path /home/fangg/tts/save_voice/tmp --out_path data_splits/voise --topk 4 --matching_layer 6 --synthesis_layer 6 Matching weightings: tensor([0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], device='cuda:0') Synthesis weightings: tensor([0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], device='cuda:0') [LIBRISPEECH] Computing folders ['train-clean-100', 'dev-clean'] Loading wavlm. WavLM-Large loaded with 315,453,120 parameters. Feature has shape: torch.Size([355, 1024])--------------------------------| 0.00% [0/1008 00:00<00:00] Done 0/1,008 Feature has shape: torch.Size([54, 1024])-------------| 0.10% [1/1008 01:02<17:28:24 train-clean-100/voise/pre/voise-pre-11911.flac] Feature has shape: torch.Size([277, 1024])------------| 0.20% [2/1008 01:56<16:17:57 train-clean-100/voise/pre/voise-pre-12097.flac] |████████████-----------------------------------------| 24.11% [243/1008 11:22<35:49
there must be something wrong
from knn-vc.
or problem
from knn-vc.
(knn) root@fangg-MS-7B78:/home/fangg/tts/knn-vc-master# python prematch_dataset.py --librispeech_path /home/fangg/tts/save_voice/tmp --out_path /home/fangg/tts/save_voice/tmp --topk 4 --matching_layer 6 --synthesis_layer 6 Matching weightings: tensor([0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], device='cuda:0') Synthesis weightings: tensor([0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], device='cuda:0') [LIBRISPEECH] Computing folders ['train-clean-100', 'dev-clean'] Loading wavlm. WavLM-Large loaded with 315,453,120 parameters. Feature has shape: torch.Size([22, 1024])---------------------------| 0.00% [0/2583 00:00<00:00] Done 0/2,583 Feature has shape: torch.Size([55, 1024])------------| 0.04% [1/2583 00:05<3:40:41 train-clean-100/018/pre/018-pre-11006.flac] Feature has shape: torch.Size([32, 1024])------------| 0.08% [2/2583 00:05<2:00:22 train-clean-100/018/pre/018-pre-11009.flac] Done 1,000/2,583█████-------------------------------| 38.71% [1000/2583 02:52<04:33 train-clean-100/074/pre/074-pre-12003.flac] Done 2,000/2,583████████████████████████------------| 77.43% [2000/2583 03:37<01:03 train-clean-100/076/pre/076-pre-12999.flac] All done!███████████████████████████████████████████| 100.00% [2583/2583 04:01<00:00 dev-clean/015/pre/015-pre-13138.flac]flac] (knn) root@fangg-MS-7B78:/home/fangg/tts/knn-vc-master# (knn) root@fangg-MS-7B78:/home/fangg/tts/knn-vc-master# (knn) root@fangg-MS-7B78:/home/fangg/tts/knn-vc-master#
But this time there is no problem, why?
from knn-vc.
Hi @fangg2000 , there was a minor typo in the prematching script which caused some unexpected behaviour (see #16 ) -- I committed a fix now for it and it should operate as expected.
Note that during prematching we do compute WavLM features, so it will use GPU memory, and the loading of files can take quite a long time if your disk is slow and dataset is massive, but it shouldn't take extremely long.
Hope that helps!
from knn-vc.
Related Issues (20)
- out_wav is a wav file? HOT 1
- Question about WavLM layer choice HOT 2
- Torch Hub CPU inference support HOT 1
- prematch argument HOT 2
- Choice for k HOT 2
- Conversion output has very strong similarity to source audio. HOT 3
- Considering context around source features HOT 2
- extend to other SSL model features
- Will this work with 44100hz audio? HOT 4
- Error in quickstart HOT 2
- Size mismatch error HOT 2
- loss issues encountered in fine-tuning the model HOT 6
- Will this work for singing voice conversion (svc)? HOT 2
- Maybe mention memory consumption in readme.md? HOT 3
- Using another batch size in training HOT 2
- Some questions about implementation HOT 2
- Matching pool empty HOT 2
- Hints on improvements for training and matching HOT 3
- How to plug-in new finetuned HiFiGAN? HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from knn-vc.