Fused Triton kernels for late-interaction (MaxSim) scoring
-
Updated
Aug 12, 2026 - Python
Fused Triton kernels for late-interaction (MaxSim) scoring
Convert Hugging Face ColBERT retrieval models (ModernBERT & BERT) to pg_colbert GGUF, and export standard llama.cpp embedding models.
To associate your repository with the pylate topic, visit your repo's landing page and select "manage topics."