Archive: 2023 May

News
A Review of SpeechT5: Introducing Google's T5 into Speech (ASR, TTS, SID, ...) Tasks
Tutorials

SpeechT5 ๋ฆฌ๋ทฐ: ์Œ์„ฑ(ASR, TTS, SID, ...) ์ž‘์—…์— Google์˜ T5 ๋„์ž…ํ•˜๊ธฐ

T5๋Š” ๋‹จ์ผ ์ธ์ฝ”๋”-๋””์ฝ”๋” ์‚ฌ์ „ ํ›ˆ๋ จ ๋ฐฉ์‹ ์•„๋ž˜์—์„œ ํ…์ŠคํŠธ-ํ…์ŠคํŠธ ์ž‘์—…์„ ํ†ตํ•ฉํ–ˆ์œผ๋ฉฐ NLP์˜ ํ•ต์‹ฌ ๋„๊ตฌ๊ฐ€ ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

5์›” 20, 2023ยท 1 min read
A Review of Deepmind's WaveNet for TTS/Audio Synthesis (Does it look like GPT to you?)
Tutorials

Deepmind์˜ WaveNet TTS/์˜ค๋””์˜ค ํ•ฉ์„ฑ ๋ฆฌ๋ทฐ (๋‹น์‹ ์—๊ฒŒ GPT์ฒ˜๋Ÿผ ๋ณด์ด๋‚˜์š”?)

Tacotron ์ด์ „์—, VALL-E ์ด์ „์—, ์‹ ๊ฒฝ ์ฝ”๋ฑ์ด ์˜ค๋””์˜ค๋ฅผ ํ† ํฐ์œผ๋กœ ๋ณ€ํ™˜ํ•˜๊ธฐ ์ด์ „์—, DeepMind์˜ WaveNet์€ ์‹ ๊ฒฝ๋ง์ด ์›๋ณธ ์˜ค๋””์˜ค ํŒŒํ˜•์„ ์ƒ˜ํ”Œ ๋‹จ์œ„๋กœ ์ƒ์„ฑํ•  ์ˆ˜ ์žˆ์Œ์„ ๋ณด์—ฌ์ค€ ๋…ผ๋ฌธ์ด์—ˆ๊ณ  ๋ชจ๋“  ๋งค๊ฐœ๋ณ€์ˆ˜์‹

5์›” 6, 2023ยท 1 min read