๋ณธ๋ฌธ ๋ฐ”๋กœ๊ฐ€๊ธฐ
TIL

[2024.01] 1์ฃผ์ฐจ Today I Learned

by rahites 2024. 1. 1.

01/01 ์›”

๐ŸŒž Happy New Year!!! ๐ŸŒž

1. ํˆฌ๋น…์Šค Seamless ๋…ผ๋ฌธ ํ™•์ธ

https://arxiv.org/abs/2312.05187

 

Seamless: Multilingual Expressive and Streaming Speech Translation

Large-scale automatic speech translation systems today lack key features that help machine-mediated communication feel seamless when compared to human-to-human dialogue. In this work, we introduce a family of models that enable end-to-end expressive and mu

arxiv.org

2. Opic ์˜์ƒ 1๊ฐœ

 

01/02 ํ™”

1. ํˆฌ๋น…์Šค Seamless ๋…ผ๋ฌธ ํ™•์ธ

 

01/03 ์ˆ˜

1. ํˆฌ๋น…์Šค Seamless ๋…ผ๋ฌธ ์ •๋ฆฌ

- SeamlessM4T v2, Seamless Expressive, Seamless Streaming, ์ฃผ์š” Metrics

 

01/04  ๋ชฉ

1. 2024๋…„ ์ฒซ ์ถœ๊ทผ

- ๋ฆฌ๋ˆ…์Šค ๊ถŒํ•œ ๊ด€๋ฆฌ(chmod ์ˆซ์ž(rwx))

- ssh, scp

 

2. AutoDub Inference

- Whisper ๋ชจ๋ธ ์‚ฌ์šฉ์‹œ ์˜ค๋ฅ˜ ๋ฐœ์ƒํ•˜์—ฌ Papago API๋กœ ์‹คํ–‰

- Papago + Whisper ๋ชจ๋ธ ์‚ฌ์šฉ์‹œ ์ด์ „๋ณด๋‹ค ๋” ์ข‹์€ ๊ฒฐ๊ณผ ํ™•์ธ

https://github.com/WiFiHan/autodub

 

GitHub - WiFiHan/autodub

Contribute to WiFiHan/autodub development by creating an account on GitHub.

github.com

 

01/05  ๊ธˆ

1. ์„œ๋ฒ„ ํ†ต์‹  ๋ฐฉ๋ฒ•

- TCP/IP, UDP ์ฐจ์ด์ 

- PORT ์‚ฌ์šฉ๋ฐฉ๋ฒ•

 

2. Service ํŒŒ์ผ ์‚ฌ์šฉ

- /etc/systemd/system/ ํด๋”๋กœ ์‹คํ–‰ํ•  service ํŒŒ์ผ ๋ณต์‚ฌ

- systemctl start ~.service๋กœ ์„œ๋น„์Šค ์‹คํ–‰

- status ๋ช…๋ น์–ด๋กœ ํ˜„์žฌ ์ƒํƒœ ํ™•์ธ ๊ฐ€๋Šฅ

- daemon ์‹คํ–‰์‹œ service ํŒŒ์ผ์˜ ๋‚ด์šฉ์„ ๋ณ€๊ฒฝํ•  ๊ฒฝ์šฐ daemon reload ํ•„์š”

- service๋ฅผ ์‹คํ–‰ํ•  ๋•Œ ๋ณ€์ˆ˜๋ฅผ ์ž…๋ ฅํ•˜์—ฌ ๋™์ผํ•œ ํŒŒ์ผ๋กœ ์—ฌ๋Ÿฌ ์„œ๋น„์Šค๋ฅผ ์‹คํ–‰ํ•  ์ˆ˜ ์žˆ๋„๋ก ๋งŒ๋“ค ์ˆ˜ ์žˆ์Œ

 

์ถ”ํ›„ ์ •๋ฆฌ ์˜ˆ์ •...

 

01/06  ํ† 

1. Vall-e-x ๋…ผ๋ฌธ ์ฝ๊ธฐ

https://arxiv.org/pdf/2303.03926v1.pdf

2. ์˜์–ด Speaking ์Šคํ„ฐ๋””

'TIL' ์นดํ…Œ๊ณ ๋ฆฌ์˜ ๋‹ค๋ฅธ ๊ธ€

[2024.01] 2์ฃผ์ฐจ Today I Learned  (0) 2024.01.08
[2023.12] 4์ฃผ์ฐจ Today I Learned  (0) 2023.12.25
[2023.12] 3์ฃผ์ฐจ Today I Learned  (0) 2023.12.19

๋Œ“๊ธ€