embedding-comparison

SKILL.md

Evaluate and compare different embedding models on your actual data.

The default all-MiniLM-L6-v2 model is a good starting point, but may not be optimal for your specific content. This skill helps you:

| Domain vocabulary | Technical jargon may need specialized models | | Document length | Some models handle long text better | | Query style | Conversational vs keyword queries | | Speed requirements | Larger models = better quality but slower | | Memory constraints | Some models need significant RAM |

의미론적 검색을 위한 임베딩 모델 비교 및 ​​평가 출처: mindmorass/reflex.

원본 보기