Ryan Prenger

Senior Deep Learning Researcher at NVIDIA

[Ryan Prenger @DeepPrenger]

Try Googling "Who played Glinda in the Wizard of Oz"

[Gavi Begtrup @GaviBegtrup]

Here's a Secret: #startups are Brutal
Please RT to spread the word on my FIRST @inc column
https://t.co/BmbxxJED4T

[Ryan Prenger @DeepPrenger]

There are also talks from HuggingFace, PayPal and more!

[Ryan Prenger @DeepPrenger]

We have a pretty awesome line up at the conversational AI track at GTC!

https://t.co/CgBIo22ttc

Tomorrow I'll be moderating talks from Facebook AI Research on dialogue agents in Minecraft as well as a talk on GPT-3 from OpenAI!

[Ryan Prenger @DeepPrenger]

I'm tearing up a little bit.

[Bryan Catanzaro @ctnzr]

We’ve had a dream for many years of a TTS engine with enough emotional range to do voice acting. We’re incredibly proud to unveil Flowtron, the TTS narrating the 2020 I AM AI video. Take a listen: https://t.co/J6jDe1TdC7

[Bryan Catanzaro @ctnzr]

We’ve had a dream for many years of a TTS engine with enough emotional range to do voice acting. We’re incredibly proud to unveil Flowtron, the TTS narrating the 2020 I AM AI video. Take a listen: https://t.co/J6jDe1TdC7

[Ryan Prenger @DeepPrenger]

https://t.co/J6ymH34j0B

[Rafael Valle @RafaelValleArt]

We finally shared a Mellotron model pre-trained on LJS for our essential artistic needs :)
https://t.co/XyzIUXcFXa

[Apollo @maoam_guy]

@angaidh_ @benrileysmith

[Rafael Valle @RafaelValleArt]

Very often I'd turn around from my desk to be delighted with blowing examples from Deep Learning Super Sampling 2.0! Definitely a game changer! DLSS 2.0 is "a model to rule them all", just like WaveGlow! Take a look and compare DLSS with other methods!

[Ryan Prenger @DeepPrenger]

@cfm_miku_en

[Rafael Valle @RafaelValleArt]

We just released the paper and code for Mellotron: a multispeaker voice synthesis model that can make a voice emote and sing without emotive or singing training data.
https://t.co/XyzIUXcFXa

[Rafael Valle @RafaelValleArt]

Celebrate WaveGlow's anniversary with us! Listen to our new samples and download our new model weights with higher audio quality and higher throughput. WaveGlow latest implementation produces audio samples at a rate of 4850 kHz! 😱 https://t.co/3LSj8khZ22

[AI at Meta @AIatMeta]

The stable release of @PyTorch 1.0 is here! A big thank you to the community that's formed around #PyTorch and to those contributing with code, feedback, or new projects. Learn how to get started here: https://t.co/GX8poSLL6o

[Bryan Catanzaro @ctnzr]

Image padding causes small but measurable artifacts for CNNs. Partial convolution padding improves ResNet-50 Top1 by 0.478% on average, and improves semantic segmentation results on image borders. Paper by @GuilinL: https://t.co/PGUQYkkKpW Code: https://t.co/XmEb0cm626

[Rafael Valle @RafaelValleArt]

We have published our Tacotron 2 and WaveGlow trained model! Visit our repository https://t.co/TWZVhS9s8c to check out our text-to-speech synthesis demo and build your own high quality TTS models with faster than real-time inference!

[Bryan Catanzaro @ctnzr]

A good explanation of how WaveGlow works. Note that our inference speed is now 1200 kHz. 🔥

[Bryan Catanzaro @ctnzr]

WaveGlow inference: now faster! We improved mel-spectrogram inversion speed from 500 kHz to 1200 kHz on a single V100 GPU. The model is the same, just implemented better. More to come! https://t.co/JICEKsd00V

[David Duvenaud @DavidDuvenaud]

We just open-sourced a suite of ODE solvers in PyTorch:
https://t.co/y3INuQUs9C
Everything happens on the GPU and is differentiable. Now you can use ODEs in your deep learning models! Credit to @rtqichen.

[Ryan Prenger @DeepPrenger]

@ml_review @RafaelValleArt @ctnzr Wanted to give a shout out to @DrNyssa for the logo idea. And to @BSozoo for the Beon font.