Notice
Recent Posts
Recent Comments
Link
| 일 | 월 | 화 | 수 | 목 | 금 | 토 |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 8 | 9 | 10 | 11 | 12 | 13 | 14 |
| 15 | 16 | 17 | 18 | 19 | 20 | 21 |
| 22 | 23 | 24 | 25 | 26 | 27 | 28 |
| 29 | 30 | 31 |
Tags
- dockerfile
- Vite
- vscode
- podman
- getting started
- path
- FileChannel
- Typescript
- io
- 시작하기
- Webpack
- Ollama
- tauri
- config
- React
- Java
- podman compose
- .dockerignore
- curl
- RandomAccessFile
- qwen3-coder-next
- File
- glm-ocr
- docker
- Docker Compose
- html canvas
- cli
- 절대경로
Archives
- Today
- Total
워로디스
ollama glm-ocr의 컨텍스트를 작게 고정해서 새 모델로 만들기 본문
glm-ocr는 128K 컨텍스트 모델로 배포됩니다.
대부분 OCR에는 128K까지 필요 없어서 16K 정도로 낮추면 로딩 문제가 사라지는 경우가 많습니다.
q8_0(메모리 절약) + 16K 컨텍스트 추천
ollama pull glm-ocr:q8_0
Modelfile 파일을 하나 만들고(현재 폴더에 생성)
FROM glm-ocr:q8_0
PARAMETER num_ctx 16384
그 다음
ollama create glm-ocr-16k -f Modelfile
ollama run glm-ocr-16k "Text Recognition: ./image.png"
glm-ocr의 기본 사용 예시는 ollama 라이브러리 문서에 있는 형태(프롬프트에 이미지 경로 포함) 그대로 쓰면 됩니다.
만약 16K로도 부족/문제면: 8192 → 16384 → 32768 순서로 올리면서(필요 최소로) 테스트하세요.
'정리 > ollama' 카테고리의 다른 글
| ollama qwen3-coder-next 메모리 사용 조정 (0) | 2026.02.20 |
|---|
