No module named 'vc_infer_pipeline'

[AbdurrehmanSubhani]

Getting the following error while using RVC:

Downloading text_2.pt: 100% 5.35G/5.35G [01:19<00:00, 67.1MB/s] Downloading (…)solve/main/vocab.txt: 100% 996k/996k [00:00<00:00, 5.19MB/s] Downloading (…)okenizer_config.json: 100% 29.0/29.0 [00:00<00:00, 2.30kB/s] Downloading (…)lve/main/config.json: 100% 625/625 [00:00<00:00, 43.2kB/s] Downloading coarse_2.pt: 100% 3.93G/3.93G [01:03<00:00, 45.2MB/s] Downloading fine_2.pt: 100% 3.74G/3.74G [01:54<00:00, 30.2MB/s] Downloading: "https://dl.fbaipublicfiles.com/encodec/v0/encodec_24khz-d7cc33bc.th" to /root/.cache/torch/hub/checkpoints/encodec_24khz-d7cc33bc.th 100%|██████████| 88.9M/88.9M [00:00<00:00, 146MB/s]

ModuleNotFoundError Traceback (most recent call last)

in () 30 31 if use_rvc: ---> 32 from rvc_infer import get_vc, vc_single 33 get_vc(rvc_path, device, is_half) 34 # simple generation /content/bark-with-voice-clone/rvc_infer.py in 111 sys.path.append(now_dir) 112 sys.path.append(os.path.join(now_dir,"Retrieval-based-Voice-Conversion-WebUI")) --> 113 from vc_infer_pipeline import VC 114 from lib.infer_pack.models import SynthesizerTrnMs256NSFsid, SynthesizerTrnMs256NSFsid_nono, SynthesizerTrnMs768NSFsid, SynthesizerTrnMs768NSFsid_nono 115 from fairseq import checkpoint_utils ModuleNotFoundError: No module named 'vc_infer_pipeline'

[dagshub[bot]]

Join the discussion on DagsHub!

[yintju03]

I have same question

[ugiatenc]

Same

[gutodiasdev]

Same

[gustrd]

This repo is not very active, this fork from BrasD99 is more updated:

https://github.com/BrasD99/bark-with-voice-clone/tree/main

[kikoferrer]

Need support for this issue. Unable to install via pip

[Furanofsky]

Same problem, I played around a bit with the https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI and the vc_infer_pipeline.py from https://github.com/Tiger14n/RVC-GUI/ or https://huggingface.co/spaces/ardha27/rvc-models/tree/main but without success. Maybe I am using a wrong path.

Edit: Fixed it, the zip "RVC-beta.7z" from https://huggingface.co/lj1995/VoiceConversionWebUI/tree/main has the correct file paths and all needed files. You can integrate it as "Retrieval-based-Voice-Conversion-WebUI" into Bark or set your own path. (I am on Windows)

[jswildone]

@Furanofsky That's the same approach I've been working toward this evening. I'm not sure if you've seen it, but I went and instead grabbed RVC1006Nvidia.7z as it is a more recent RVC update with some really useful new features for training/inference, as opposed to RVC-beta.7z. I think 1006 allows for way smoother and faster training, if you would be interested in the change. I'm still troubleshooting the install but think the hassle of installing it now will be better in the long run compared to the betas based on the changelog.

I actually ended up running into the same problem as you and OP with the: ModuleNotFoundError: No module named 'vc_infer_pipeline' error and just thought to ask how you went about solving it? You mentioned how you integrated RVC in Bark, which I interpret as you making a folder (within the Bark directory) titled: 'Retrieval-based-Voice-Conversion-WebUI' but just thought to clarify if that was what ended up solving it for you in the end or if you ended up doing more than that?

I noticed when looking through the ipynb's that it specified: rvc_path = f"Retrieval-based-Voice-Conversion-WebUI/weights/{rvc_name}.pth"

But with 1006, there doesn't seem to be a weights folder directly under the main folder. I instead found one at this directory instead: D:\Programming\Bark\Retrieval-based-Voice-Conversion-WebUI\assets\weights

Just thought to throw a bit of context, should it help if you go on to install 1006 and run into a similar problem.