fituna good first issues (Python)

tests: cover every OS branch of _detect_ram_mb() in hardware.py / hardware.py의 _detect_ram_mb() OS별 분기 테스트 추가 2 days ago

good first issue area: platform

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

tests: add pytest coverage for the GGUF header parser in model_info.py / model_info.py의 GGUF 헤더 파서에 pytest 테스트 추가 2 days ago

good first issue area: search

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

Edge devices (Raspberry Pi, Jetson): benchmark on the device without running the full search there / 엣지 기기(Raspberry Pi, Jetson): 전체 탐색 없이 기기에서 벤치마크만 실행 4 days ago

enhancement help wanted area: platform roadmap

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

AMD GPU auto-detection unsupported on Windows (no rocm-smi) - investigate alternative detection paths / Windows에서 AMD GPU 자동 감지 미지원(rocm-smi 없음): 대체 감지 방법 조사 4 days ago

enhancement help wanted area: platform

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

Post-recommendation integration: launch the config, or export it to Ollama/LM Studio / 추천 이후 연동: 구성을 바로 실행하거나 Ollama/LM Studio로 내보내기 4 days ago

enhancement help wanted roadmap

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

Multi-GPU support: --tensor-split / 멀티 GPU 지원: --tensor-split 4 days ago

enhancement help wanted area: platform roadmap

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

docs: make the issue and PR templates bilingual (English + Korean) / 이슈·PR 템플릿을 영어·한국어 병기로 바꾸기 4 days ago

documentation good first issue area: docs

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

Call for test reports: run FiTuna on your hardware (Linux, Windows, Raspberry Pi, and more) / 다양한 환경에서의 실행 결과와 버그 제보를 받습니다 4 days ago

help wanted good first issue area: platform

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

English translations for docs/ technical documents 13 days ago

documentation good first issue area: docs

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

Follow-up to #74: skip the redundant baseline cache write, stop touching cache internals in tests / #74 후속: 불필요한 baseline 캐시 쓰기 제거, 테스트의 캐시 내부 접근 정리 17 days ago

good first issue area: search

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

docs: reflect the single-pass KLD baseline (#74) in ARCHITECTURE / ARCHITECTURE 문서에 #74 변경(KLD 기준값 1회 측정) 반영 17 days ago

documentation good first issue area: docs

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

Native-speaker review of English docs and code comments / 영문 문서·주석을 자연스러운 영어로 다듬기 17 days ago

documentation help wanted good first issue area: docs

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

README roadmap: link every item to its tracking issue / README 로드맵 항목마다 이슈 링크 추가 18 days ago

documentation good first issue area: docs

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

No pytest coverage for the llama.cpp output parsers 20 days ago

help wanted good first issue area: search

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

KLD mode: reuse the baseline PPL from the logits-generation run instead of a separate pass 23 days ago

enhancement good first issue area: search

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

Add KL divergence as an optional quality metric about 1 month ago

enhancement good first issue measurement-integrity

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

CODE_OF_CONDUCT.md에 비공개 신고 연락처 명시 about 2 months ago

documentation good first issue

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

doctor: exit code 2 is reported when a binary check crashes, not only when a binary is missing 3 months ago

bug good first issue

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

Real-hardware validation on Windows 3 months ago

help wanted area: platform

leeyunseokarchive/fituna

0

Stop guessing your llama.cpp config. Measure it — hardware-benchmark-driven auto-tuning for local LLMs.

Python

leeyunseokarchive/fituna

Open Issues Need Help