20b
PulseAugur coverage of 20b — every cluster mentioning 20b across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
Local LLM Memory Needs: Weights, KV Cache, and Usable RAM Explained
Estimating the memory required for local LLMs involves calculating the sum of model weights, the KV cache, and overhead, then comparing this to the usable memory available on a device. Quantization significantly impacts…
-
New benchmark tests LLMs on Danish cultural heritage
Researchers have developed SDUs DAISY, a new benchmark designed to evaluate large language models' understanding of Danish cultural heritage. The benchmark, derived from the Danish Culture Canon 2006 and Wikipedia, feat…
-
User trains 1.1B LLM from scratch for $200, shares code and model
A user has successfully trained a 1.1 billion parameter large language model from scratch for approximately $200. The model, named 'gemmeh', was pre-trained on 20 billion tokens from the fineweb-edu dataset and then fin…
-
20B MoE AI Model Runs at 120 Tokens/Sec on iPhone
A new demonstration, Maple-Preview, showcases a 20-billion parameter Mixture-of-Experts (MoE) model capable of running at 120 tokens per second on an iPhone. This development highlights the increasing efficiency and on-…