🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

New blog: New Years resolutions for PyTorch in 2025

Zizheng was one of our interns at NVIDIA back in summer 2023. Later, when we were considering to make him a FT offer, he chose to join DeepSeek without much hesitance. Back then, the DeepSeek multimodal team only has 3 people. I am still very much impressed by Zizheng’s decision

Anyone know of a 2025 version of this article on how to select a GPU workstation in 2025? Or have advice? The main thing I'm uncertain of is consumer cards (5090) vs workstation cards (A6000) and how to weigh differences in VRAM, tensor cores, etc. timdettmers.com/2023/01/30/whi…

As we are learning DeepSeek is one of the most sophisticated psyops of all time. Here's how it went down: 1) Release the model open source. 2) Include highly detailed papers for all other people to replicate your work. 3) Create a novel SOTA RL algo that uses less memory

🚀 Big news for the Cloud Native Geospatial community! Apache Iceberg and #Parquet now support native geometry and geography (GEO) data types, thanks to grassroots efforts from Planet, CARTO, Wherobots, and many others. Huge thanks to the Iceberg community for collaborating with

I'm a big believer in all the AI advances but I also think there can be alot of hype sometimes. I think it's important to share null experiments as well as successes to understand where things are at. Sometimes people only share positive experiments and hold back null

Just show this video to every lawmaker in the country.

Very excited to have you on the team Isaac!

Today is World Autism Awareness Day. My wife and I want to raise awareness of and help decrease any stigma around autism by sharing that my wife learned she has autism later in life—in her late 30s. It’s common for girls and women to have autism overlooked due to their often

Inspired by SamGEO, we've enabled mapping objects at scale, predicting boundaries or bounding boxes with simple text prompts. Check out the blog to learn how we made this happen and how you can try it in SQL!

If you’re at CNG this week, come chat with Ryan Avery and I about how to combine TorchGeo models + the MLM STAC extension to create performant and scalable geospatial ML inference pipelines!

#geospatial foundation models aren't scaling well, so we should probably stop training and scaling them the way we currently are. In this blogpost I sit down and crunch the numbers to show this: bit.ly/4kycY4H