Massimo Morelli
@massimomorelli
Ex business consultant, now business intelligence consultant. Works for Data Reply in Munich. Mostly moved to bsky.app/profile
ID: 2879881
http://it.linkedin.com/in/ingmorelli 29-03-2007 19:05:15
6,6K Tweet
965 Takipçi
572 Takip Edilen
🚀🚀 How does the Apache Iceberg reading process look like? ◉ The reader first visits the catalog to retrieve the table's current metadata file location. ◉ After fetching the metadata file, it collects the table’s schema and checks partition schemes to understand the data
Remember the llm.c repro of the GPT-2 (124M) training run? It took 45 min on 8xH100. Since then, Keller Jordan (and by now many others) have iterated on that extensively in the new modded-nanogpt repo that achieves the same result, now in only 5 min! Love this repo 👏 600 LOC
On the heels of Humanity's Last Exam, Scale AI & Center for AI Safety have released a new very-hard reasoning eval: EnigmaEval: 1,184 multimodal puzzles so hard they take groups of humans many hours to days to solve. All top models score 0 on the Hard set, and <10% on the Normal set 🧵