{"id":23985549,"url":"https://github.com/codename0og/codename-rvc-fork","last_synced_at":"2025-04-14T09:51:34.199Z","repository":{"id":239059472,"uuid":"798387518","full_name":"codename0og/codename-rvc-fork","owner":"codename0og","description":"Retrieval-based-Voice-Conversion ( RVC ) modified and enhanced by codename;0","archived":false,"fork":false,"pushed_at":"2024-07-08T22:38:31.000Z","size":793,"stargazers_count":8,"open_issues_count":0,"forks_count":2,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-03-27T23:11:12.695Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/codename0og.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-05-09T17:14:53.000Z","updated_at":"2025-02-18T16:43:55.000Z","dependencies_parsed_at":"2024-06-18T11:43:45.098Z","dependency_job_id":"c8b6e77e-b3e6-4d86-8005-8676495b8521","html_url":"https://github.com/codename0og/codename-rvc-fork","commit_stats":null,"previous_names":["codename0og/codename-rvc-fork"],"tags_count":1,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/codename0og%2Fcodename-rvc-fork","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/codename0og%2Fcodename-rvc-fork/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/codename0og%2Fcodename-rvc-fork/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/codename0og%2Fcodename-rvc-fork/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/codename0og","download_url":"https://codeload.github.com/codename0og/codename-rvc-fork/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248859649,"owners_count":21173337,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-01-07T14:16:26.203Z","updated_at":"2025-04-14T09:51:34.172Z","avatar_url":"https://github.com/codename0og.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"## Essentially an 'up-to-date RVC' \u003cbr /\u003e  \u003cbr /\u003e✨ Enhanced with everything I find useful, practical and neat ✨ \u003cbr /\u003e  \u003cbr /\u003eㅤㅤㅤFork creator:   Codename;0 \u003cbr /\u003eㅤ\n### **If you need to contact me, here's my discord:ㅤ.codename.0.**\n\u003e tho, please, let me know right away you come from github.\n\u003e Also, keep in mind I don't keep people on my friend list for too long,\u003cbr /\u003eso don't get surprised with the removal after we sort our things out lol - no hate, that's how I am.\n\u003cbr /\u003e\n\n#### ㅤ \u003cbr /\u003eImportant info:\n\u003e 1. This fork is using mainline RVC ( repo, not releases ) as a base.\n\u003e 2. Crucial point of this fork is to be up-to-date with the mainline.\n\u003e 3. Some features and ui elements ( overal ui style ) borrowed from mangio's fork. - I kinda dislike the new one lol.\n\u003e 4. I might port / borrow few things from other forks 'cause why not. If there are thigs **you** want to see added in here, let me know.\n⠀\u003cbr /\u003e\n⠀\u003cbr /\u003e\n## ⚠️ PYTHON COMPATIBILITY WARNING ⚠️\n### The fork is confirmed and tested-in-field to work with python 3.10.6 with my requirements list. \u003cbr /\u003e( DO NOT USE python 3.12.0, it won't work. )\n\n\n\u003cbr /\u003e\n\n\n# Fork's features:\n**| Training related |**\n\n- Mangio's crepe F0/Feature extraction method\nIt is the best ( Not the fastest tho. ) extraction method for models that rely on clean ( No reverb, delay, harmonies, noise etc. ) or truly HQ datasets.\n- Adjustable hop_length for Mangio's crepe method ( Definitely the biggest perk of using Mangio-crepe. )\n- Envelopes for processed samples / segments to avoid zero-crossing ( waveform interruption ) clicks.\n- My own \" Mel similarity metric \" as a bonus. Helps in spotting overtraining / overfitting and mode collapses. Metric is being displayed in the console / log and is also logged in tensorboard files.\n- Sox is used for resampling ( vhq - Very High quality as default )\n\n⠀\u003cbr /\u003e\n**| Inference related |**\n\n- Automatic index file matching for models.\n- Auto detection of audio files in \" audios \" folder + a dropdown to choose 'em.\n- Changed default Feature index search ratio to '0.5'.\n- Mangio-crepe ( standard; \" full \" ) - Go to method.\n- FCPE f0/pitch estimation method added - tweaked and ported from applio by me.\n\n⠀\u003cbr /\u003e\n**| ONNX Export related |**\n- There's been many issues and problems with newest releases' onnx exporting so I decided to port 0618's ( RVC 0618 v2 beta release ) way of handling it.\n( tl;dr, now exporting works. It exports the onnx models with onnx 14's operators.  ~ Tested having onnx 15 runtime - so the latest one available. )\n\nI've confirmed the exported .onnx models to work with:\n- my \"RVC_Onnx_Infer\" thingy: https://github.com/codename0og/RVC_Onnx_Infer ( and generally rvc's own original onnx inference demo ) \n- As for w-okada, nope. Shapes mismatch.\n( I suppose that's pretty normal and in order to use w-okada with onnx, you gotta convert the rvc's .pth to .onnx in the w-okada itself. )\n\n⠀\u003cbr /\u003e\n## Bugs / Stuff that's broken.\n\n- Noticed that after you unload the model from inference the \" protect voiceless consonants \" value becomes null.\nFIX: After you unload the voice / model and then load any again, simply move the slider or input the default value: 0.33\n( I've re-checked the code again during the big update and.. honestly? idk, code seems fine and handling too so idk why it happens lol.\nIt's either something that's always been there and I didn't notice or mangio-style ui breaks somethin' I dunno. whatever. )\n\n- No other bugs or issues I've observed. In case you do and it's serious / critical, write me a msg on discord.\n\n⠀\u003cbr /\u003e\n## Potentially to-do ( might be abandoned if I change my mind lel ):\n- onnx inference within the web-ui ( separate tab ) + have it use rvc's slicing / caching mechanism⠀\u003cbr /\u003e(( I have it in wip state in older dev fork build, the ui and onnx model loading in works, but it's all just a dummy input. No inference yet\n- voice changer within the web-ui\n- re-organize all the files ( structure re-interpretation\n- change / tweak the style of the interface\n- model similarity comparitor ( haven't gotten into it yet )\n- automated 'accurate' log interval syncing\n\n⠀\u003cbr /\u003e\n## CONCEPT STAGE - Stuff I kinda have in mind but am unsure if I wanna go that way.\n- Native ONNX inference\nI kinda worked on it but mehh, due to the reasons I've written about in my onnx infer thingy I sorta don't feel a need to work on it, at least for now.\n\n- new additional slicing mechanism.. perhaps whisper or maybe other ml assisted approach ( would have a checkbox )\n- incorporating yt-dlp and ffmpeg conversion for yt-sourced audio.. hmm.. maybe even all in 1 'simple' dataset preparator lol, maybe.\n- in-built tts solution\n- auto dl'er for voice models ( dunno of that yet )\n- dataset evaluation mechanism ( fidelity scoring )\n\u003cbr /\u003e\n\n\n# Installation and usage guide:\n### Everything is easy and automated. Just read carefully and follow please:\n\u003cbr /\u003e\n\n- Step 1. Download the \"codename-rvc-fork-v1.0.0.zip\" from the release section.\n\u003cbr /\u003e\n\n- Step 2. Unpack the content of the .zip ( 'codename-rvc-fork-v1.0.0' folder ) to your preferred location.\n\n#### ⚠️ IMPORTANT: ㅤThe path to the unpacked rvc folder cannot contain spaces or funky symbols.\n#### An example of a good path:\n\u003e D:\\AI_related\\voice-cloning\\codename-rvc-fork-v1.0.0\n\u003cbr /\u003e\nㅤ\nㅤ\n\n- Step 3. In the fork's folder, go to:ㅤㅤ'requirements/RVC' and pick the right .bat for you:\n\n#### For standard Nvidia users ( GTX 10xx and RTX users ) :\n\u003e ' install requirements for RVC (  CUDA117  ).bat '\n\n#### IF THE ABOVE ↑ CAUSES YOUR GPU TO BE UNDETECTED ( some RTX 40xx cases), \u003cbr /\u003eUSE THE FOLLOWING ONE:\n\u003e ' install requirements for RVC (  CUDA118  ).bat '\n\u003cbr /\u003e\n\nㅤ\nㅤ\n- Step 4. Go back to the main rvc folder and use this to download required assets:\n\u003e ' download-required-assets.bat '\n\u003cbr /\u003e\n\n- Step 5. Run the RVC with:\n\u003e ' RUN_WEB-UI.bat '\n\u003cbr /\u003e\n\n\n### Where to put downloaded RVC voice models?\n\u003e Put the '.pth' model files in:ㅤㅤ'assets/weights'\n\u003cbr /\u003e\n\n\n\u003e Put the index files:ㅤㅤ'logs/folder_with_the_same_name_as_the_pth_model'\n\u003cbr /\u003e\n\n### END OF THE GUIDE\n\u003cbr /\u003e\n\n# Here are my other thingies if you're interested:\n- Standalone and lightweight RVC's native real-time-voice-changer:\n\u003e https://github.com/codename0og/rvc-realtime-voice-changer\n- Inferencing for RVC's onnx models:\n\u003e https://github.com/codename0og/RVC_Onnx_Infer\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcodename0og%2Fcodename-rvc-fork","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fcodename0og%2Fcodename-rvc-fork","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcodename0og%2Fcodename-rvc-fork/lists"}