Twitter | Search | |
Loreto Parisi
Computer Engineer MSc. Technical Director of Machine Learning Father of an amazing girl and a cute boy.
12,320
Tweets
710
Following
938
Followers
Tweets
Loreto Parisi retweeted
Pieter Levels ✨ May 23
Replying to @Patreon
Also considering doing some kind of fund so I can help creatives for 12 months with for ex $1k - $2k/mo to feed themselves, and work on launching lots of projects to see which one sticks.
Reply Retweet Like
Loreto Parisi May 24
Replying to @rocksetta
nooo I was thinking to do exactly the same thing for my son! How did you control the step motors, any raspberry / arduino + BT/Wifi or a magic wand🧐?
Reply Retweet Like
Loreto Parisi May 24
...where the NBTHK here Swords catalogue has an annotated dataset of swords labeling, that's amazing!
Reply Retweet Like
Loreto Parisi May 24
Replying to @baykenney @braddwyer
Did you tried cartographic aerial photos?
Reply Retweet Like
Loreto Parisi May 24
Replying to @tkasasagi
It would be the first (names) dataset ever!
Reply Retweet Like
Loreto Parisi retweeted
Alfredo Canziani May 22
Replying to @PyTorch
Prediction and Policy-learning Under Uncertainty (PPUU) Slides: Article: code: Website: Poster:
Reply Retweet Like
Loreto Parisi May 23
This answer is far better than the question!
Reply Retweet Like
Loreto Parisi May 23
concurrent vs asynchronous requests in
Reply Retweet Like
Loreto Parisi May 23
Longest Increasing Subsequence of Indexes in a Array
Reply Retweet Like
Loreto Parisi May 23
Reply Retweet Like
Loreto Parisi May 23
Replying to @tkasasagi
And...you were not using time machine
Reply Retweet Like
Loreto Parisi May 22
Too much confidence in systems, whatever advanced driver assistance system they are, can be fatal. It’s clear that the use of these systems must be regulated and not put in production as-it-is.
Reply Retweet Like
Loreto Parisi May 22
Replying to @citnaj @pjreddie
Agreed. When moving from images to movies with darkflow ( darknet + tensorflow) I actually find that was simpler workin on frame changes (obtained with opencv) while keeping the same model.
Reply Retweet Like
Loreto Parisi May 22
My name is on the next mission to Mars: NASA's Rover. Send your name too!
Reply Retweet Like
Loreto Parisi May 22
Replying to @citnaj
Great! Do you mean that you should add a specific 3D layer to handle consecutive frames?
Reply Retweet Like
Loreto Parisi retweeted
Olivier Bousquet May 20
We recently launched a joint postdoc program between and Google Research Zurich! If you are interested in this unique opportunity of getting both academic and industrial research experience, apply now! via
Reply Retweet Like
Loreto Parisi retweeted
roadrunner01 May 20
Few-Shot Adversarial Learning of Realistic Neural Talking Head Models pdf: abs:
Reply Retweet Like
Loreto Parisi May 21
Reply Retweet Like
Loreto Parisi retweeted
Uber Engineering May 19
Horovod is an open source distributed deep learning system that enables to scale model training from one to eight GPUs for their self-driving sensing and perception technologies.
Reply Retweet Like
Loreto Parisi May 18
Replying to @citnaj
That’s impressive! Which is the maximum video frame rate supported in real-time?
Reply Retweet Like