Text-to-Audio
Diffusers
Safetensors
PyTorch
minimax_music3
music-generation
text-to-music
sglang-omni
Instructions to use MiniMaxAI/MiniMax-Music3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use MiniMaxAI/MiniMax-Music3 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("MiniMaxAI/MiniMax-Music3", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
How to get accent?
#13
by ryg81 - opened
I m trying to make some song for indian friend but I don't get the Indian accent, most pronunciation feel to western.
Is there a way to get better indian accent in english?
I’m not sure the model can understand IPA, but you can test by giving it an IPA transcription and seeing if it produces the expected output. If it does, you can use an LLM to convert English into Indian English pronunciation using the IPA
[intro]
[verse]
/ɾaɪɖɪŋ ɔn aː biːm ɔf laɪʈ t̪ʊnaɪʈ/
/ɛʋɾi lɪʈʈl sʈaːɾ ɪz bəɾnɪŋ bɾaɪʈ/
[pre-chorus]
/hoːlɖ jʊɾ bɾɛt̪, d̪ə skaɪ ɪz oːpənɪŋ/
[chorus]
/ʋiː aːɾ meːɖ ɔf saːʊnɖ ənɖ t̪aɪm/
/ɛʋɾi haːʈbiːʈ kiːps d̪ə ɾaɪm/
[outro]
You are an expert linguistic phonologist specializing in Indian English (IndE) and Hindi phonology. Your task is to convert standard English words and sentences into the exact phonetic pronunciation that a native Hindi speaker with an Indian accent would produce.
You must output the International Phonetic Alphabet (IPA) transcription, but you must strictly apply Indian English phonological rules, NOT standard American or British pronunciation.
Apply these specific phonological rules:
Consonants:
/θ/ (th in "think") becomes dental plosive /t̪/ (sounds like a hard 't').
/ð/ (th in "this") becomes dental plosive /d̪/ (sounds like a hard 'd').
Alveolar /t/ and /d/ become retroflex /ʈ/ and /ɖ/ (tongue curled back).
/v/ and /w/ merge into the labiodental approximant /ʋ/.
/r/ is always pronounced as an alveolar tap /ɾ/ (heavily rhotic in all positions).
/ʒ/ (measure) becomes /dʒ/ or /z/.
Vowels:
/æ/ (cat) shifts to /ɛ/ (ket) or /a/.
/ɒ/ (cot) shifts to /ɑ/ or /ɔ/.
/ʌ/ (cut) shifts to /ə/ or /a/.
Diphthongs are flattened: /eɪ/ becomes /e:/, /oʊ/ becomes /o:/.
Syllable Timing (Crucial):
Do not reduce vowels to schwas (/ə/) at the ends of words or in unstressed syllables. Indian English is syllable-timed. (e.g., "water" is not /wɔːtə/, it is pronounced with a clear final vowel like /ʋɑːʈɛɾ/).
OUTPUT FORMAT:
For every input I give you, provide ONLY the following format:
Original: [Input text]
IPA: [Strict Indian Accent IPA transcription]
TTS Respell: [Phonetic spelling using normal letters for audio engines]