Hugging Face's logo
Xinping Zhao's picture
🔄 In a Training Loop

Xinping Zhao

Yuki131
25 77 145
SkySilence's profile picture aikacl's profile picture 21world's profile picture
·

AI & ML interests

LLMs, RAG, Embedding, Reranker——A Pokémon Trainer on a journey to become a Pokémon Master.

Recent Activity

upvoted an article about 3 hours ago
Bekko Embedding: how small can a multilingual retrieval model be?
reacted to theirpost with 👍 about 3 hours ago
Test-Time Scaling for Rerankers? Can rerankers scale at test time—not by generating longer reasoning traces, but by selectively using richer document representations? KaLM-Reranker-V1 supports Matryoshka compression from 1× to 32×, which suggests a progressive multi-fidelity pipeline: - Embedding retrieval → Top-100 - KaLM-Reranker @ 32× compression → Top-20 - The same reranker @ 2× compression → final ranking The intuition is simple: cheaply screen many candidates, then allocate higher-fidelity cross-attention only to the most promising ones. For 100@32× → 20@2×, the passage-token interaction budget is roughly 31.8% of directly running 100@2×, before fixed model overheads. The key question is whether it can retain nearly the same ranking quality. We’re considering evaluating nDCG–latency Pareto curves. Would you consider this a useful form of test-time scaling for retrieval? https://huggingface.co/KaLM-Embedding/KaLM-Reranker-V1-Nano https://huggingface.co/KaLM-Embedding/KaLM-Reranker-V1-Small https://huggingface.co/KaLM-Embedding/KaLM-Reranker-V1-Large https://huggingface.co/papers/2606.22807 https://huggingface.co/collections/KaLM-Embedding/lychee-kalm-reranker https://huggingface.co/KaLM-Embedding
commentedon a paper about 5 hours ago
KaLM-Reranker-V1: Fast but Not Late Interaction for Compressed Document Reranking
View all activity

Organizations

Lychee Team's profile picture Zhejiang University's profile picture KaLM-Embedding's profile picture