SECourses LeVo Song Generator - This is the Very Best local Song generator APP - I can say almost level of Suno AI Music - We have so many batch feature
Patreon exclusive posts index to find our scripts easily, Patreon exclusive posts index to see which scripts updated or added last and amazing Patreon special generative scripts list that you can use in any of your task.
Join discord to get help, chat, discuss and also tell me your discord username to get your special rank : SECourses Discord
Please also Star, Watch and Fork our Stable Diffusion & Generative AI GitHub repository and join our Reddit subreddit and follow me on LinkedIn (my real profile)
=======
Latest installers zip file : Levo_Song_Gen_v12.zip
App Screenshots Gallery (Public) :
https://www.reddit.com/r/SECourses/comments/1r9g2d8/secourses_levo_song_generator_this_is_the_very/
App Example Generation and Features :
The app is almost perfect please report errors you see
We have batch processing, number of generations, full preset saving and loading, 20 pre-made presets, generating a song with each preset and many other features
Official repo : https://github.com/tencent-ailab/SongGeneration
Input audio for consistent voice is limited to 10 seconds because the model works like an LLM and has fixed context size therefore input audio is using a big chunk of available fixed context size (i think context size is 8k)
Windows Requirements
Python 3.10, FFmpeg, CUDA 12.8, C++ tools, MSVC and Git
If it doesn't work make sure to below tutorial and install everything exactly as shown in this below tutorial
15 March 2026 V11.0
This is another massive update
We have fixed the issue with uploading reference voice
Song Structure Tags tab and automatic check of tags sysytem updated for the latest version
The model will obey 100% to your set Duration (seconds)
If your lyrics is not enough you can later cut your generated song where it ends
Lesser the song duration, it will use lesser VRAM
New Auto VRAM Set Optimize added
I have added Enable LLM MLP-Only Int8 and Enable LLM MLP-Only Int4
I coded this and this is first in the community for this model
Int8 is recommended
With Int8 activation now 180 seconds song takes as low as 12 GB
Most of the VRAM is used at kv-cache since it uses auto-regressive model even i make model int8 or int4, it doesnt make massive difference in peak vram usage
Moreover new block swapping implemented to advanced settings
Since an LLM is used for generating song tokens in this model architecture, block swapping works really really slow but if you get OOM, you can enable it
Most of the VRAM is used at kv-cache since it uses auto-regressive model and block swapping really can't fix it
Example Lyrics updated and the video posted above is generated with that one with default settings except i made duration like 220 seconds
Some various bugs fixed and improvements made
For updating / installing, download latest v11 zip file, overwrite all previous files or extract into a new folder, and run Windows_Install_and_Update.bat and wait for entire install to be finished
It will download all necessary models as well automatically
Then use Windows_Start_App.bat run
13 March 2026 V10.1
VRAM management improved and made even better
There was a new published model songgeneration_v2_large 10 March 2026 and now our default model is this and previous models all removed
The progress displayed on the CMD even further improved
Now Generation Type works accurately mixed, vocal only, bgm only or both seperate
Latest changes from official repo implemented in our app and now our app has all features and more
I recommend make a fresh install with latest V10 zip file
Install into a new folder, verify and then delete older installation
20 Feburary 2026 V9.0
Installers are all upgraded to latest version and now we use uv installation
Thus the installations are much faster like 10x to 100x literally
Several bugs fixed and app is working perfect and amazing
I recommend to make a fresh install with latest zip file and delete older directory completely
9 November 2025 V7.0
App 100% upgraded to newest models and newest libraries
Make a fresh install
Download SongGeneration Large - BEST QUALITY model
So far this model tested and working
Hopefully I will do more test and optimizations
Now can generate up to 4.5 minute songs with amazing quality
It will auto download SongGeneration Large - BEST QUALITY model
To download other model run Windows_Download_or_Resume_Models.bat
4 August 2025 V4
Audio input was not working and causing stuck in model loading and now fixed
Moreover i modified the app so that now you can both upload audio and video and it will auto extract audio from your videos
Windows_Resume_Download_Models.bat file added to resume failed downloads
New feature Extra Description (Optional) added
You can change tempo of your song etc with this like pop, uplifting, the bpm is 130
Moreover if you enable Use only extra prompt it will skip using Genre, Instrument, Emotion, Timbre, Gender dropdown boxes so you have full control
Also BPM slider added to the app
Furthermore, it will display final Description on cmd like : Description: male, bright, rock, uplifting, electric guitar and drums, the bpm is 135
Generation Type moved to the main tab,
Generate vocals+BGM (mixed), vocals only, or BGM only
So you can generate only vocals or only background music or both (default)
Default preset is updated to produce better results and still researching better generation workflow, best to test more
I think write your song only with [verse] and [chorus], dont use [pre-chorus] or [post-chorus]
I think [drop], [hook], [instrumental] can be also useful
All included in the new default preset, i also updated the video here with generated first sample
Model is also able to generate 2.5 minute songs
Just run Windows_Install_and_Update and it should work
Massed Compute (Recommend Cloud) :
Please register via this link : https://vm.massedcompute.com/signup?linkId=lp_034338&sourceId=secourses&tenantId=massed-compute
Use our coupon SECourses
Our coupon works on all GPUs now
H100 has amazing price and speed but you can use like RTX A6000 ADA as well
Full details here : https://www.patreon.com/posts/26671823
Then select our image SECourses from Creator dropdown
Then follow Massed_Compute_Instructions_READ.txt
Same as my any other Massed Compute installer script
Example tutorial for learn how to install and use Massed Compute
(Starts at 12:58) : https://youtu.be/KW-MHmoNcqo?si=G1WbG-Qw4ujWvOtG&t=778
RunPod (Cloud):
Please register via this link : https://get.runpod.io/955rkuppqv4h
Then follow Runpod_Instructions_READ.txt
Same as my any other RunPod installer script
Use the template written in Runpod_SimplePod_Levo_Song_Instructions.txt file
Example tutorial for learn how to install and use RunPod
(starts at 22:03) : https://youtu.be/KW-MHmoNcqo?si=QN8X8Sjn13ZYu-EU&t=1323