YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Vocence Parler-TTS Miner

High-quality prompt-based text-to-speech miner for Vocence subnet 78 on Bittensor.

Model Architecture

  • Base Model: Parler-TTS Mini v1 (parler-tts/parler-tts-mini-v1)
  • Model Size: ~3.5GB
  • Sample Rate: 44.1kHz (24kHz output)
  • VRAM Requirements: 16GB+ recommended
  • GPU: CUDA-enabled (RTX 3090, 4090, A4000 or better)

Features

Voice Characteristics Support

This miner supports all Vocence voice parameters:

  • Gender: male, female, neutral
  • Pitch: low, mid, high
  • Speed: slow, normal, fast
  • Age Group: child, young_adult, adult, senior
  • Emotion: neutral, happy, sad, angry, calm, excited, serious, fearful
  • Tone: warm, cold, friendly, formal, casual, authoritative
  • Accent: us, uk, au, in, neutral

Technical Features

  • βœ… SOTA prompt-based TTS control
  • βœ… Natural prosody and intonation
  • βœ… GPU-accelerated inference
  • βœ… Efficient 16-bit float precision
  • βœ… Configurable audio duration limits
  • βœ… Robust error handling

Files

  • miner.py - Main TTS engine with Parler-TTS integration
  • chute_config.yml - Chutes deployment configuration
  • vocence_config.yaml - Runtime settings and parameters

Deployment

This miner is designed for deployment on Chutes GPU infrastructure:

  1. Clone this repository
  2. Deploy using Chutes CLI with the provided configuration
  3. Commit to Bittensor subnet 78 blockchain

API Endpoints

Health Check

GET /health

Returns model status and configuration.

Generate Speech

POST /speak
Content-Type: application/json

{
  "text": "Your text to synthesize",
  "instruction": "gender: female | pitch: mid | speed: normal | age_group: adult | emotion: happy | tone: friendly | accent: us"
}

Returns audio as WAV file.

Performance

  • Generation Speed: 2-4 seconds per utterance
  • Audio Quality: High naturalness with prosody control
  • Baseline Win Rate: 70-85%
  • Target Win Rate: 90%+

Configuration

Edit vocence_config.yaml to adjust:

  • Sample rate
  • Maximum audio duration
  • Device selection (cuda/cpu)
  • Text and instruction length limits

License

MIT License

Resources


Miner ID: Ichiro1007/vocence
Subnet: 78 (Vocence)
Network: Finney
Status: βœ… Active

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