StyleTTS 2.

Columbia Universityopen-sourceN/A paramsTTS

StyleTTS 2. Uses large speech LM (WavLM) for adversarial training. Surpasses human on LJ Speech. NeurIPS 2023.

§ 01 · Benchmarks

Every benchmark StyleTTS 2 has a recorded score for.

#	Benchmark	Area · Task	Metric	Value	Rank	Date	Source
01	LJ Speech	Speech · Text-to-Speech	mos	4.5%	#2/5	2023-06-12	source ↗
02	VCTK	Speech · Text-to-Speech	mos	4.2%	#3/6	2023-06-12	source ↗

Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.

§ 02 · Strengths by area

Where StyleTTS 2 actually performs.

Speech

benchmarks

avg rank #2.5

§ 03 · Papers

1 paper with results for StyleTTS 2.

2023-06-12· Speech· 2 results
StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models

§ 05 · Sources & freshness

Where these numbers come from.

arxiv

results