VoxCPM2 Fine-tuned for Indic Languages
Fine-tuned from openbmb/VoxCPM2 on Indic language TTS data.
Languages
- Hindi
Training Details
- Mode: lora (LoRA r=32)
- Epochs: 5
- Learning Rate: 0.0001
- Datasets: SPRINGLab/IndicVoices-R_Hindi
Usage
from voxcpm import VoxCPM
model = VoxCPM.from_pretrained(
"chandrashekars/VoxCPM2-Indic-LoRA",
load_denoiser=False,
)
# Generate Hindi speech
wav = model.generate(
text="नमस्ते, यह VoxCPM2 है।",
cfg_value=2.0,
inference_timesteps=10,
)
import soundfile as sf
sf.write("output.wav", wav, model.tts_model.sample_rate)
- Downloads last month
- 39
Model tree for chandrashekars/VoxCPM2-Indic-LoRA
Base model
openbmb/VoxCPM2