Google Brain team researchers @ekindogus, @barret_zoph, and @quocleix have open sourced the AutoAugment image classification policies that achieve state-of-the-art results. Check out the the paper and code is located in tensorflow/models/research/autoaugment. #AutoML https://t.co/p4ZNxbzh5t
— Google AI (@GoogleAI) September 27, 2018
Happy to announce the release of Scikit-learn 0.20.0! It's available on pypi and conda now! Check the release notes here: https://t.co/r8os0f03EK I'm excited about the ColumnTransformer! This release is dedicated to the memory of Raghav Rajagopalan. Thanks for everybody's help!
— Andreas Mueller (@amuellerml) September 26, 2018
I would like to announce that after a few months of hard work we (@albuslaev @creaf @cvtalks @viglovikov) are releasing image augmentation library #albumentations v0.1.1 with an extended bounding box support and many other improvements. https://t.co/TditWXpg8J pic.twitter.com/h2jU7xOkap
— Vladimir Iglovikov (@viglovikov) September 26, 2018
apricot: Smart, #sklearn compatible subsampling for #DataScience & #MachineLearning. #Python
— Randy Olson (@randal_olson) September 25, 2018
Sometimes datasets contain a lot of redundant information & you're wasting time training on unnecessarily large datasets. This library will help eliminate that.https://t.co/v784J5bl0d pic.twitter.com/tG0YoZI9FJ
More than you'd ever wanted to know about the choices I've made in gganimate #rstats https://t.co/JjPKW2PC09
— Thomas Lin Pedersen (@thomasp85) September 24, 2018
Data Scribers: A simple and efficient way to browse data science blogs https://t.co/SVDXimEvWS pic.twitter.com/Cgr3snYPzh
— Mike Kearney📊 (@kearneymw) September 23, 2018
mlcrate 0.2 is out! ✨
— Mikel Bober-Irizar (@mikb0b) September 23, 2018
Highlights:
- New rank_average() ensembling method
- Apache Feather support for save/load
- Simple no-fuss conversion between NumPy/Python types and @PyTorch tensors with tonp() and totensor()https://t.co/J81CKF5D7t / pip install mlcrate pic.twitter.com/9CWkObHFFF
✨Today @Uber announced an open-source project called Petastorm, which enables single-machine or distributed training + evaluation of deep learning models directly from datasets in @ApacheParquet format.@TensorFlow, @ApacheSpark, @PyTorch are supported!https://t.co/Gjv4X0Ww5A pic.twitter.com/9TzDK9wlBR
— 👩💻 @DynamicWebPaige ✨ (@DynamicWebPaige) September 22, 2018
💞 so many little details to love in {radix}
— Mara Averick (@dataandme) September 21, 2018
👉 "Radix for R Markdown: Scientific and technical writing, native to the web" by @fly_upside_down, @riannone, & @xieyihuihttps://t.co/kzAWcA78E1 #rstats pic.twitter.com/3nswE17e52
And I thought stable baselines in RL were an oxymoron ... https://t.co/BOn62MAPD7
— hardmaru (@hardmaru) September 20, 2018
Announcing Radix, a new R Markdown format for web-native scientific and technical writing: https://t.co/mVSkWlSMjA #rstats #rmarkdown pic.twitter.com/fUzYJL1w45
— RStudio (@rstudio) September 19, 2018
🐝 Bees! Bees everywhere! Run for your lives!
— Mara Averick (@dataandme) September 19, 2018
📊 "ggbeeswarm: Column-scatter / beeswarm-style plots in ggplot2" by @pleiotropehttps://t.co/XraUkE9axc #rstats #dataviz pic.twitter.com/1JCXAShhUo
