Suresh (@_Suresh2)온디바이스 LoRA는 여전히 전체 베이스 모델을 로드해야 하며, 어댑터가 KV 캐시 크기도 줄여주지 않는다는 기술적 한계를 지적한다. 경량 추론과 메모리 최적화 측면에서 중요한 개발 이슈다.https://x.com/_Suresh2/status/2049275115158815208#lora #ondevice #llm #kvcache #inference
Related
AI in Cybersecurity, Part 32: Automating Active Directory Pentests with ARESARES - an open-core autonomous red-team plat...
AI in Cybersecurity, Part 32: Automating Active Directory Pentests with ARESARES - an open-core autonomous red-team platform that orchestrates 66+ attack modules, maps enterprise t...
Wake me up when AI can...https://stoppels.ch/2026/10/03/wake-me-up-when-ai-can.html#ai
Wake me up when AI can...https://stoppels.ch/2026/10/03/wake-me-up-when-ai-can.html#ai
https://omdaruliterature.blogspot.com/2026/11/cassiopeia-wiki-selected-where-to-start.html Cassiopeia - Wiki - Selected ...
https://omdaruliterature.blogspot.com/2026/11/cassiopeia-wiki-selected-where-to-start.html Cassiopeia - Wiki - Selected - Where to Start - AInquiries #AI #Essays in #English on #sp...