Text-to-Audio
MLX
Safetensors
apple-silicon
singing-voice-synthesis
singing-voice-conversion
soulx-singer
Instructions to use mlx-community/SoulX-Singer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/SoulX-Singer with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir SoulX-Singer mlx-community/SoulX-Singer
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
File size: 579 Bytes
62c5237 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 | infer:
n_steps: 32
cfg: 3
audio:
hop_size: 480
sample_rate: 24000
max_length: 36000
n_fft: 1920
num_mels: 128
win_size: 1920
fmin: 0
fmax: 12000
mel_var: 8.14
mel_mean: -4.92
model:
encoder:
vocab_size: 3000
text_dim: 512
pitch_dim: 512
type_dim: 512
f0_bin: 361
f0_dim: 512
num_layers: 4
flow_matching:
mel_dim: 128
hidden_size: 1024
num_layers: 22
num_heads: 16
cfg_drop_prob: 0.2
use_embedding: False
cond_codebook_size: 512
cond_scale_factor: 1
sigma: 1e-5
time_scheduler: cos |