<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
<title>CanRun 가이드</title>
<link>https://canrunllm.com/ko/guides</link>
<description>내 PC나 맥에서 로컬 LLM을 돌리는 방법을 다룬 가이드입니다.</description>
<language>ko</language>
<lastBuildDate>Wed, 30 Sep 2026 00:00:00 GMT</lastBuildDate>
<atom:link href="https://canrunllm.com/ko/rss.xml" rel="self" type="application/rss+xml"/>
<item><title>로컬 LLM 사양 — VRAM 8~32GB로 돌릴 수 있는 모델 (2026)</title><link>https://canrunllm.com/ko/guides/vram-requirements</link><guid isPermaLink="true">https://canrunllm.com/ko/guides/vram-requirements</guid><description>그래픽카드 VRAM 8·12·16·24·32GB에서 어떤 모델이 돌아가는지 정리했습니다. 필요한 메모리를 계산하는 법과, 컨텍스트·양자화에 따라 달라지는 점도 함께 다룹니다.</description><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate></item>
<item><title>Ollama가 느리거나 앞 내용을 잊을 때 — 컨텍스트 기본값</title><link>https://canrunllm.com/ko/guides/ollama-context-length</link><guid isPermaLink="true">https://canrunllm.com/ko/guides/ollama-context-length</guid><description>Ollama가 긴 대화의 앞부분을 버리거나 느려지는 이유를 설명합니다. VRAM별 기본 컨텍스트 길이와 늘리는 방법, 늘렸을 때 더 드는 메모리도 정리했습니다.</description><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate></item>
<item><title>GGUF 양자화 등급 정리 — Q4_K_M·IQ4_XS·Q8_0 무엇을 받을까</title><link>https://canrunllm.com/ko/guides/gguf-quantization</link><guid isPermaLink="true">https://canrunllm.com/ko/guides/gguf-quantization</guid><description>GGUF 양자화 단계마다 파일 용량이 얼마인지, 품질은 어떻게 나뉘는지 정리했습니다. 한 단계만 낮춰도 못 돌리던 모델이 돌아가는 경우를 데이터로 보여 줍니다.</description><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate></item>
<item><title>한국 오픈 LLM 정리 — EXAONE·Kanana·HyperCLOVA X·Solar 구동 사양</title><link>https://canrunllm.com/ko/guides/korean-open-llms</link><guid isPermaLink="true">https://canrunllm.com/ko/guides/korean-open-llms</guid><description>LG EXAONE, 카카오 Kanana, 네이버 HyperCLOVA X SEED, 업스테이지 Solar를 내 PC에서 돌리려면 무엇이 필요한지 정리했습니다. 모델 사양과 라이선스, 흔한 GPU·맥에서의 판정을 함께 봅니다.</description><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate></item>
<item><title>맥에서 로컬 LLM 돌리기 — M4·M5 통합 메모리는 얼마나 필요할까</title><link>https://canrunllm.com/ko/guides/mac-local-llm</link><guid isPermaLink="true">https://canrunllm.com/ko/guides/mac-local-llm</guid><description>맥의 통합 메모리 가운데 모델이 실제로 쓸 수 있는 양과 M4·M5 구성별로 돌아가는 모델을 정리했습니다. 속도를 좌우하는 메모리 대역폭도 함께 다룹니다.</description><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate></item>
<item><title>KV 캐시와 컨텍스트 길이 — 컨텍스트를 늘리면 VRAM이 얼마나 더 필요할까</title><link>https://canrunllm.com/ko/guides/kv-cache-context</link><guid isPermaLink="true">https://canrunllm.com/ko/guides/kv-cache-context</guid><description>컨텍스트가 길어지면 왜 메모리가 더 드는지, 모델마다 KV 캐시가 토큰당 얼마나 차지하는지 정리했습니다. KV 캐시를 q8_0·q4_0으로 양자화하면 얼마나 줄어드는지도 다룹니다.</description><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate></item>
</channel>
</rss>
