Last 90 days · 2608 total
An independent researcher developed IMGNet, a face verification model that achieves 96.27% accuracy on LFW by using sliding window sign pattern match…
OriginalA Reddit thread invites community members to discuss and share their favorite local Visual Language Models (VLMs), along with detailed reasons and ev…
OriginalA Hacker News "Show HN" post demonstrates how to successfully run the GLM 5.2 model on a computer with limited resources.
OriginalThe EU Parliament has given its approval to Chat Control 1.0, a measure aimed at regulating digital communications.
OriginalA Reddit post discusses successfully running the massive GLM-5.2 (744B MoE) model on a consumer machine with only 25GB of RAM.
OriginalAgentLens is introduced as a new production-assessed benchmark for evaluating interactive code agents through detailed trajectory reviews, moving bey…
OriginalNew research explores integrating Large Language Model (LLM)-powered reasoning into agent-based modeling (ABM) to enhance the capabilities of modelin…
OriginalTriRoute introduces a unified learned routing mechanism for jointly adapting attention, managing Mixture-of-Experts (MoE) allocations, and optimizing…
OriginalResearch proposes using a self-critical masked language model for generating effective ad headlines, addressing the challenge of creating compelling…
OriginalNew research applies fine-tuned latent diffusion models to synthesize novel motifs for Ulos textiles, an Indonesian cultural heritage, balancing pres…
Original