Hugging Face's logo
Activity Feed

AI & ML interests

Building the foundation for better search with embeddings, rerankers, RAG infrastructure, and small language models.

Recent Activity
cosyy  updated a model about 14 hours ago
KaLM-Embedding/KaLM-Reranker-V1-Nano
cosyy  new activity about 14 hours ago
KaLM-Embedding/KaLM-Reranker-V1-Large:fix(reranker): avoid re-computing the first batch in predict()'s batch-size probe to reduce additional computational effort
cosyy  new activity about 14 hours ago
KaLM-Embedding/KaLM-Reranker-V1-Nano:fix(reranker): avoid re-computing the first batch in predict()'s batch-size probe to reduce additional computational effort
View all activity

Papers

KaLM-Reranker-V1: Fast but Not Late Interaction for Compressed Document Reranking

LMEB: Long-horizon Memory Embedding Benchmark

View all Papers

Organization Card
Community About org cards

KaLM-Embedding: Building the foundation for better search with embeddings, rerankers, RAG infrastructure, and small language models.

Lychee-KaLM develops embedding models, rerankers, RAG infra, and small language models for better search. We aim to develop efficient and accessible systems that help people discover, retrieve, and use knowledge more effectively.

Explore
🌐 KaLM-Embedding Homepage · 💻 GitHub · 🗂️ Embedding Training Data · 📖 Text Embedding Survey · 📚 LMEB Benchmark · 🏆 LMEB Leaderboard · 🔎 KaLM-Embedding Demo · ⚡ KaLM-Reranker Demo

Collections 6

View 6 collections