Twitter | Search | |
Jason Antic
Software guy, currently digging deep into GANs to do some cool photo colorization and restoration! Author of DeOldify.
916
Tweets
1,969
Following
1,278
Followers
Tweets
Jason Antic retweeted
Jeremy Howard 14h
This is the most thoughtful analysis of the model release issue I've seen. I'm not sure I agree with it. But I'm sure of this: those NLP academics who responded with shallow snarky toxic uninformed envy are an embarrassment to the field.
Reply Retweet Like
Jason Antic retweeted
helena sarin Feb 16
Replying to @glagolista
Hiems Operimentum is propagated by SNGAN exclusively early experiments show its great potential as all unnatural floor covering
Reply Retweet Like
Jason Antic retweeted
Dmitri Petrov Feb 15
Reply Retweet Like
Jason Antic retweeted
Michael E. Mann Feb 16
Replying to @rstackjr @NaomiOreskes
Please read what has already said. Good scientists aren't ALARMISTS. Our message may be--and in fact IS--ALARMING. The distinction is so very, very critical and cannot be brushed under the rug.
Reply Retweet Like
Jason Antic retweeted
d00d Feb 16
Me: "Can you tell me a joke?" 's GPT-2 117M: "Don't waste your life in conversation Silence is a fundamental living principle. It's a natural defense system, designed to give us refuge from the pain of socialized misery and pain of hyper-masculinity."
Reply Retweet Like
Jason Antic retweeted
Colorize Images Feb 16
Colorize your black & white photos automatically for free:
Reply Retweet Like
Jason Antic retweeted
Garry Tan Feb 16
Harvard MBA: That's the perfect résumé. Buffett: Well when are you going to start doing what you like? MBA: I'll get to that someday. Buffett: Your plan sounds to me a lot like saving up sex for your old age. It just doesn't make a lot of sense.
Reply Retweet Like
Jason Antic retweeted
Bored Yann LeCun Feb 15
I am actually an Artificial General Intelligence developed by ... but he kept my source code secret because if it got out and BYLs proliferated, too many fools would get .
Reply Retweet Like
Jason Antic Feb 16
Wonderful advice.
Reply Retweet Like
Jason Antic retweeted
Bored Yann LeCun Feb 15
Just trained a ConvNet so deep it gets 200% ImageNet accuracy. BUT, the weights are so big, in your clown novices hands they would form a singularity & devour the planet. So I'm gonna keep that shit locked up & just Tweet/blog about it.
Reply Retweet Like
Jason Antic retweeted
hardmaru Feb 13
Pretty cool how StyleGAN can generate entire fonts in one shot.
Reply Retweet Like
Jason Antic retweeted
Law Boy, Esq. Feb 13
excited to be entering my second decade of using the term "Kafkaesque" without ever having read Kafka. i will never stop and i will never be caught.
Reply Retweet Like
Jason Antic Feb 13
Ohhh I like this one.
Reply Retweet Like
Jason Antic Feb 11
This was a fun interview! Sanyam let me say exactly what I wanted to say.
Reply Retweet Like
Jason Antic Feb 11
I really like Daniel Dennett and his work but I’ve always thought that he basically just skirted around the “hard problem” of consciousness and never really had a satisfying answer for it....
Reply Retweet Like
Jason Antic retweeted
helena sarin Feb 11
Replying to @glagolista
north american evergreen orchid, deer resistant...
Reply Retweet Like
Jason Antic Feb 11
This is a great example of doing the much needed work of “refactoring deep learning”. I love how elegant this is!
Reply Retweet Like
Jason Antic Feb 11
Very interesting to hear. People are the most important part of images to get right I think. Thanks for sharing this!
Reply Retweet Like
Jason Antic retweeted
colourisesg Feb 11
We used ’s architecture and trained it on Singaporean historical images and Open Images. And so we plan to release our model some time in March, in addition to a CoLab notebook that you can use to colourise your own photos on Google Drive.
Reply Retweet Like
Jason Antic Feb 11
Great stuff- what an honor! I am curious- how much of a difference did Open Images make? I've been considering it but it's huuuuuuuge.
Reply Retweet Like