Codesota · Models · VALL-E 2Microsoft2 results · 2 benchmarks
Model card

VALL-E 2.

MicrosoftproprietaryUnknown paramsNeural codec language model (EnCodec tokens)1 current SOTA

VALL-E 2. First system achieving human parity on LibriSpeech. Grouped code modeling + repetition aware sampling. Jun 2024.

§ 02 · Benchmarks

Every benchmark VALL-E 2 has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01LJ SpeechAudio · Text-to-speechmos4.6%#1/52024-06-08source ↗
02VCTKAudio · Text-to-speechmos4.2%#4/62024-06-08source ↗
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.
§ 03 · Strengths by area

Where VALL-E 2 actually performs.

Audio
2
benchmarks
avg rank #2.5 · 1 SOTA
§ 04 · Papers

1 paper with results for VALL-E 2.

  1. 2024-06-08· Speech· 2 results

    VALL-E 2: Neural Codec Language Models are Human Parity Zero-Shot Text to Speech Synthesizers

§ 05 · Related models

Other Microsoft models scored on Codesota.

Phi-4
14B params · 17 results
Phi-4 Multimodal Instruct
6B params · 9 results
WizardLM-2-8x22b
7 results
DeBERTa-v3-large
304M params · 6 results
E5-Mistral-7B-instruct
7B params · 4 results
ResNet-152
60M params · 3 results
ResNet-50
25M params · 3 results
UniXcoder
Unknown params · 3 results
§ 06 · Sources & freshness

Where these numbers come from.

arxiv
2
results
2 of 2 rows marked verified.