You can deploy this model and start earning money today!

MiniMaxMusicTraining/
soad-mm3-nextlat-xm-daron-contrand30s-20260824-5k-adamw2e-5-singersplit-reginst-capdrop100

Name: soad-mm3-nextlat-xm-daron-contrand30s-20260824-5k-adamw2e-5-singersplit-reginst-capdrop100
Availability: InStock

diffusers

minimaxmusic

minimaxmusic-diffusers

text-to-audio

audio

simpletuner

not-for-all-audiences

lora

sd-lora

standard

MiniMaxAI/MiniMax-Music3

adapter

apache-2.0

Views

No views yet

MiniMaxMusicTraining/soad-mm3-nextlat-xm-daron-contrand30s-20260824-5k-adamw2e-5-singersplit-reginst-capdrop100

This is a PEFT LoRA derived from MiniMaxAI/MiniMax-Music3.

Validation was disabled during training.

The text encoder was not trained. You may reuse the base model text encoder for inference.

Training settings

Training epochs: 416
Training steps: 5000
Learning rate: 2e-05
- Learning rate schedule: constant_with_warmup
- Warmup steps: 50
Max grad value: 0.5
Effective batch size: 1
- Micro-batch size: 1
- Gradient accumulation steps: 1
- Number of GPUs: 1
Gradient checkpointing: True
Prediction type: autoregressive_next_token
Optimizer: adamw_bf16
Trainable parameter precision: Pure BF16
Base model precision: no_change
Caption dropout probability: 100.0%
LoRA Rank: 64
LoRA Alpha: None
LoRA Dropout: 0.1
LoRA initialisation style: default
LoRA mode: Standard

Training modes

MiniMax Music train component: language_model (global LM / RVQ planner)
MiniMax Music LM max frames: 128
NextLat: Enabled
- Block index: -1
- Weight: 0.1
- State loss: smooth_l1
- KL weight: 0.0
XM: Enabled
- Candidate count: 2
- Selection scope: block
- Training target: route
- Block size: 16

Datasets

soad-daron-vocals

Repeats: 0
Total number of audio files: 6
Duration buckets: audio (6)
Used for regularisation data: No

soad-daron-instrumental-regularisation

Repeats: 0
Total number of audio files: 6
Duration buckets: audio (6)
Used for regularisation data: Yes

Inference

python

import torch
from diffusers import DiffusionPipeline

model_id = 'MiniMaxAI/MiniMax-Music3'
adapter_id = 'MiniMaxMusicTraining/soad-mm3-nextlat-xm-daron-contrand30s-20260824-5k-adamw2e-5-singersplit-reginst-capdrop100'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "An astronaut is riding a horse through the jungles of Thailand."
negative_prompt = 'blurry, cropped, ugly'

## Optional: quantise the model to save on vram.
## Note: The model was not quantised during training, so it is not necessary to quantise it during inference time.
#from optimum.quanto import quantize, freeze, qint8
#quantize(pipeline.transformer, weights=qint8)
#freeze(pipeline.transformer)
    
pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu') # the pipeline is already in its target precision level
model_output = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    num_inference_steps=30,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(42),
    width=256,
    height=256,
    guidance_scale=7.5,
).images[0]

model_output.save("output.png", format="PNG")

Continuation crop and regularisation tournament

This gallery compares Daron-only, NextLat + Explorative Modeling, random continuation crops capped at 30.72 seconds, instrumental regularisation, 100% caption dropout checkpoints under 5 caption conditions. Every render uses the same unseen lyrics with every vocal section assigned to the target singer, 300 seconds maximum duration, 30 flow steps, flow CFG 1.7, and seed 20260824. Within each batch, only LoRA strength changes.

The adapters were trained with 100% caption dropout.

The caption controls separate trigger response from descriptive style learning and unrelated-style leakage:

Off-genre, no trigger: Alternative pop-jazz, 102 BPM, brushed drums, upright bass, clean electric piano, airy male vocals, smooth verse-chorus songwriting, restrained dynamics, polished intimate mix.
Off-genre + trigger: Alternative pop-jazz, 102 BPM, brushed drums, upright bass, clean electric piano, airy male vocals, smooth verse-chorus songwriting, restrained dynamics, polished intimate mix, system of a down, soad style music.
Trigger only: system of a down, soad style music
Descriptors, no trigger: Lilting hypnotic arpeggios, a gentle melodic mid-tempo groove with layered harmonies, sudden explosive distorted riffs, unstable meter changes, theatrical male vocals, tightly syncopated bass and drums, abrupt shifts between fragile verses and frantic heavy choruses.
Trigger + descriptors: system of a down, soad style music, lilting hypnotic arpeggios, a gentle melodic mid-tempo groove with layered harmonies, sudden explosive distorted riffs, unstable meter changes, theatrical male vocals, tightly syncopated bass and drums, abrupt shifts between fragile verses and frantic heavy choruses.

Machine-readable manifest

checkpoint-1000

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

checkpoint-1500

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

checkpoint-2000

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

checkpoint-2500

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

checkpoint-3000

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

checkpoint-3500

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

checkpoint-4000

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

checkpoint-4500

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

checkpoint-5000

strength	Off-genre, no trigger	Off-genre + trigger	Trigger only	Descriptors, no trigger	Trigger + descriptors
0.25
0.50
0.75
1.00

Singer-assigned lyrics used for every render

text

[Intro]
Count the teeth in the traffic light
Paint the silence ultraviolet

[Verse 1: Daron Malakian]
Paper crowns on the factory floor
Everybody wants a little more
Wind-up prophets in a checkout line
Selling borrowed seconds back as time

[Pre-Chorus: Daron Malakian]
Turn the dial, divide the sky
Ask the wires to tell us why

[Chorus: Daron Malakian]
Wake up the walls, let the ceiling shake
We built a feast from a plastic lake
Hands in the air while the engines stall
Who signed the order to swallow it all?

[Verse 2: Daron Malakian]
Velvet sirens in the avenue
Paint every warning a pleasant blue
I fed my shadow to the voting machine
It came back polished, quiet, and clean

[Bridge: Daron Malakian]
Softly, softly, count to four
Kick down the arithmetic door
One for the hunger, two for the show
Three for the ones who already know

[Final Chorus: Daron Malakian]
Wake up the walls, let the ceiling shake
We built a feast from a plastic lake
Hands in the air while the engines stall
Who signed the order to swallow it all?

[Outro]
Count the teeth
Kill the light
Carry the silence
Into the night

NextLat predictor and XM route-embedding sidecars are training-only and are intentionally not used during inference.