Supertonic-2 FP16 (Float16 Quantized)

Float16 quantized version of Supertone/supertonic-2. ~50% smaller than the original fp32 model.

Directory Structure

β”œβ”€β”€ onnx/
β”‚   β”œβ”€β”€ duration_predictor.onnx (0.8MB)
β”‚   β”œβ”€β”€ text_encoder.onnx (13MB)
β”‚   β”œβ”€β”€ vector_estimator.onnx (63MB)
β”‚   β”œβ”€β”€ vocoder.onnx (48MB)
β”‚   β”œβ”€β”€ tts.json
β”‚   └── unicode_indexer.json
└── voice_styles/
    β”œβ”€β”€ F1.json ~ F5.json (Female voices)
    └── M1.json ~ M5.json (Male voices)

Usage

# Clone
git clone https://huggingface.co/likelion-axp/supertonic-2-fp16

Changes from Original

  • Converted from float32 to float16 using onnxconverter-common
  • No other modifications to model architecture or weights

License

This model is distributed under the OpenRAIL-M License, same as the original.

Attribution

Supported Languages

  • Korean (ko)
  • English (en)
  • Spanish (es)
  • Portuguese (pt)
  • French (fr)
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support