BAEM1N/snowflake-arctic-embed-l-v2.0-ko-GGUF

snowflake-arctic-embed-l-v2.0-ko GGUF

GGUF conversion of dragonkue/snowflake-arctic-embed-l-v2.0-ko.

#1 on Korean Embedding Leaderboard (NDCG@5,10 avg: 84.77)

Files

Usage with llama.cpp

llama-server -m snowflake-arctic-embed-l-v2.0-ko-Q8_0.gguf --port 9020 --embedding --pooling mean

Model Info

• —Architecture: XLM-RoBERTa (24 layers, hidden=1024)
• —Parameters: ~335M
• —Embedding dimension: 1024
• —Max sequence length: 8192
• —Korean fine-tuned version of snowflake-arctic-embed-l-v2.0