Tag
Charlie Marsh reports that a campaign with 5.6 Sol reduced ty's retained memory by 38% across ecosystem projects while improving performance, emphasizing that every change was closely human-reviewed.
ReFreeKV introduces a threshold-free approach to KV cache compression that adaptively allocates budgets, eliminating the need for input-specific thresholds while maintaining full-cache performance across diverse datasets and model sizes.
Developer @charliermarsh reports reducing ty's retained memory by another 15% using Codex, achieving a total ~25% memory reduction via /goal, and expresses enthusiasm for GPT models.
The author describes replacing a 3 GB SQLite database with a 10 MB Finite State Transducer (FST) binary to optimize a Finnish-English dictionary tool, achieving a 300x reduction in memory usage while maintaining performance.