@log_pie
yxwang.io
Joined August 2010
  • Tweets46
  • Following1.7K
  • Followers195
synthetic biology will be the next next big thing. will revisit in 10 years.
1
2
Realtime on-device audio synthesis. Audio content creation is big on TikTok. As always, more coming soon.
Hanoi Hantrakul@yaboihanoi17 Dec 2021
Realtime, low-latency and on-device ML audio synthesis! These TikToks using our brand new voice change effects made my day 🤣 Inference runs realtime across all Android and iOS devices both old and new. Can't wait to see more, 2Million views and counting! 🗣️➡️🐈 & 🐮➡️🎵 MEOW!
7
Yuxuan Wang retweeted
Disney+
@DisneyPlus12 Nov 2021
*yells in Chewbacca* 📣👂 New Text-to-Speech voices ft. some of your favorite characters are available now on @TikTok! To unlock them, discover the mystery keywords. #DisneyPlusDay

TikTok Text-to-Speech Voices | Disney+ Day | Disney+

72
353
141
3,221
Anthony Levandowski to Larry Page: Google's self-driving project is broken January 9, 2016
14
137
56
1,438
Music and sound are essential to TikTok's endless creativity. After our text-to-speech feature, here's another example of how we use audio technologies to inspire creativity and help users make cool content. ⁣ Come join us! We are hiring!
TechCrunch
@TechCrunch7 Apr 2021
As competition heats up, TikTok announces six new interactive music effects for creators tcrn.ch/3t0fB4T by @sarahintampa
5
awesome and congrats!! should make a demo of "Today is Monday!". 😜 @rustyryan
RJ Skerry-Ryan
@rustyryan10 Nov 2020
New work with Ron Weiss, @EricBattenberg, @sorooshooryad, @dpkingma -- finally achieving what @log_pie and I set out to do in 2016 before switching to spectrograms: direct waveform generation from characters. (1/7) abs: arxiv.org/abs/2011.03568 samples: google.github.io/tacotron/pu…

Wave-Tacotron: Spectrogram-free end-to-end text-to-speech synthesis

We describe a sequence-to-sequence neural network which directly generates speech waveforms from text inputs. The architecture extends the Tacotron model by incorporating a normalizing flow into...

arxiv.org
4
time flies. It's true.😆
Yuxuan Wang@log_pie16 Sep 2011
i hope we are really undergoing the second neural net renaissance.
9
RT There's an E-6B Mercury off the east coast near DC. I looked because I would expect them to pop up if he tests positive. It's a message to the small group of adversaries with SLBMs and ICBMs.
709
5,930
4,291
19,704
Immigration has contributed immensely to America’s economic success, making it a global leader in tech, and also Google the company it is today. Disappointed by today’s proclamation - we’ll continue to stand with immigrants and work to expand opportunity for all.
1,138
9,299
718
62,486
I'm excited to share Capacitron, the Tacotron team's most recent contribution to the world of expressive end-to-end speech synthesis (e.g., transfer and control of prosody and speaking style). arxiv.org/abs/1906.03402
2
26
4
76
We released a new large-scale corpus of English speech derived for TTS; LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speech Dataset: openslr.org/60/ Paper: arxiv.org/abs/1904.02882
5
149
5
352
出た arxiv.org/abs/1904.02882

LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speech

This paper introduces a new speech corpus called "LibriTTS" designed for text-to-speech use. It is derived from the original audio and text materials of the LibriSpeech corpus, which has been used...

arxiv.org
9
1
58
あいかわらずリズムの生成の実験中. 強弱(Veclocity)や微妙な各打点のタイミングのゆれも含めて、Variational Autoencoderで生成するようなモデルを作ったらだいぶカッコよくなってきた気がする.
1
16
110
go Daisy! @daisystanton
IEEE SLT2018@SLT201818 Dec 2018
Daisy Stanton, @Google on "Predicting expressive speaking style from text in end-to-end speech synthesis" presenting now at Kallirhoe Hall on the first day of #SLT2018
1
PCEN rocks.
Justin Salamon@justin_salamon29 Oct 2018
PCEN is an excellent audio frontend for sound recognition in far-field recordings. But... why? And how to configure it for your application? Answers in our new paper led by @lostanlen in collab @CornellBirds @nyuMARL @NYU_CUSP #nocmig justinsalamon.com/news/per-c…
2
TikTok is a flagship product from ByteDance.
Connie Chan 🍒@conniechan4 Dec 2018
New blog post — We're only starting to see the beginning of Consumer AI apps, including TikTok and more, all of which I go into here: a16z.com/2018/12/03/when-ai-…

When AI is the Product: The Rise of AI-Based Consumer Apps | Andreessen Horowitz

There’s a new era of AI consumer-based apps spreading around the world, though starting from China. TikTok, a short-form mobile video app, was downloaded on Apple’s App store more than 104 million...

a16z.com
1
Our work on mixing input types for controlling TTS synthesis is now up on arxiv! arxiv.org/abs/1811.07240 / samples s3.amazonaws.com/representat…. See the sampling code github.com/kastnerkyle/repre…, or jump straight to the colab to try on your own colab.research.google.com/gi…
3
62
1
166
Magenta Studio - Ableton Live Plugin buff.ly/2Tx1qTD Googleの音楽生成関連のモデルのパッケージ Magentaが Ableton Live用のプラグインとしてリリースされてますね. ドラムシーケンスを生成したり、二つのシーケンスの間を補完したり... いろいろ手軽に遊べます.
1
32
2
116
友人のMaxがWaveNet/VAEで生成したサンプルのみで作ったドラムンベーストラック。ディープラーニング x サウンドデザイン. buff.ly/2AK8CEq こういう使い方はぐっときますね! プロジェクトの詳細はこちら NeuralFunk - Combining Deep Learning with Sound Design buff.ly/2OP2WBJ
2
30
1
108
Yuxuan Wang retweeted
arxiv@arxiv_org18 Oct 2018
Hierarchical Generative Modeling for Controllable Speech Synthesis. arxiv.org/abs/1810.07217
21
54
Load more