iot2 / optimization_report.md
memevis's picture
Upload folder using huggingface_hub
1c3349c verified

TTS Model Optimization Report

Overview

  • Optimization Date: 2025-05-23 06:15:14
  • Iterations Performed: 3
  • Total Strategies Tested: 32
  • Best Overall Score: 0.8722

Performance Metrics

  • Mel Cepstral Distortion: 0.9410
  • Word Error Rate: 0.0683
  • Naturalness: 0.9428
  • Intelligibility: 0.9607
  • Speaker Similarity: 0.9582
  • Prosody: 0.9428
  • Overall Quality: 0.9857

Optimization Insights

Most Effective Parameter Settings

  • attention_scale: 1.3000
  • output_scale: 1.4000
  • projection_scale: 1.6000
  • encoder_scale: 1.5000
  • decoder_scale: 1.2000
  • base_enhancement: 0.0031
  • importance_factor: 1.5597

Optimization Journey

Iteration 1: Best strategy 'intelligibility_focus_var1.5' - Score 0.8696

Iteration 2: Best strategy 'intelligibility_focus_var1.5_iter2_noise_scale_1.2' - Score 0.8664

Iteration 3: Best strategy 'intelligibility_focus_var1.5_iter3_base_enhancement_1.2' - Score 0.8722