Unwrap @RAPIDSai release 0.17 from our family to yours: #GPU-Accelerated TreeSHAP in @XGBoostProject. Decimal Types in cuDF. Datetime in cuXFilter. A podcast. And much more! Thanks for your support. Happy New Year! https://t.co/a9ZPIPAtjq
— RAPIDS AI (@RAPIDSai) December 14, 2020
Time that fathers and mothers spend with their children, select countries (1965-2010). #parenting #dataviz
— Randy Olson (@randal_olson) December 14, 2020
Source: https://t.co/12FvF2K5S1 pic.twitter.com/8Eank5zhH9
joblib 1.0.0 is out on PyPI and conda-forge: https://t.co/24jkFl6Pha
— Olivier Grisel (@ogrisel) December 14, 2020
This release changes how joblib.hash handles numpy.dtype which is necessary to support the upcoming numpy 1.20.
This changes the hash values and therefore invalidates any existing joblib.Memory cache.
Rather than a longer Twitter thread, here's an article in @Nature worth reading about “the ethical questions that haunt facial-recognition research” which discusses some of these issues in depth, with several front line AI researchers with different views.https://t.co/MYanMg5E7T
— hardmaru (@hardmaru) December 14, 2020
People are now creating fake accounts with GAN-generated pics to harass under-represented minority groups on Twitter. https://t.co/ZkvMf8HT0Y
— Jeremy Howard (@jeremyphoward) December 14, 2020
The magic of Python coroutines - an event frequency timer in just two lines of code! pic.twitter.com/m1x3aHeyRG
— Jeremy Howard (@jeremyphoward) December 13, 2020
If you missed it, our JAX/Cloud TPU talk is now up!
— James Bradbury (@jekbradbury) December 11, 2020
We announced a new way to access Cloud TPUs, allowing direct SSH access and custom code on the TPU hosts, and gave FOUR demos showing how this supercharges JAX!
Video: https://t.co/dk4CchuXxo
Slides: https://t.co/GzwwG80biA
Limited compute and memory make it challenging to deploy state-of-the-art computer vision architectures on edge devices. Researchers introduce RNNPool, a pooling operator that can enable sophisticated vision intelligence in 200KB of RAM: https://t.co/af8Myi4fpR #NeurIPS2020
— Microsoft Research (@MSFTResearch) December 11, 2020
🥳 We now have Tsinghua's CPM, a Chinese GPT-3-like model with 2.6B parameters available from 🤗 model hub! 来试试吧！https://t.co/jnw0VH3mqG@Tsinghua_Uni @zibuyu9
— Hugging Face (@huggingface) December 11, 2020
DataCamp had a cool product and great R instructors producing material. Their CEO did something awful, and the community turned away from them. But, we would have turned back if they had done ANYTHING REASONABLE to make it right. Instead, they do the wrong thing over and over. https://t.co/mbs2kIKQ1g
— Amelia McNamara (@AmeliaMN) December 10, 2020
Talking to Internet companies in China gives me the impression that their MLOps infra is away ahead.
— Chip Huyen (@chipro) December 10, 2020
Many US Internet co. are still wondering if there's value in real-time ML. Chinese counterparts are already doing real-time inference + online training.
Are my samples biased?
Some visuals stick more than others. This one visualizes the death toll of COVID19 in the US by stacking up virtual coffins behind the White House. Source: https://t.co/tf9XDhqMuN pic.twitter.com/uIJTn6zy05
— Simon Kuestenmacher (@simongerman600) December 9, 2020
