gardel/lightrag
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Use direct embedding_func from hashing_kv (do not by pass maxiumu async control)

Browse Source
This commit is contained in:
yangdx
2025-02-01 22:12:45 +08:00
parent c98a675b6c
commit c3942077a9
1 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
lightrag/utils.py
View File

@@ -511,10 +511,10 @@ async def handle_cache(
quantized = min_val = max_val = None
if is_embedding_cache_enabled:
# Use embedding cache
embedding_model_func = hashing_kv.global_config["embedding_func"]["func"]
llm_model_func = hashing_kv.global_config.get("llm_model_func")
current_embedding = await embedding_model_func([prompt])
current_embedding = await hashing_kv.embedding_func([prompt])
llm_model_func = (
hashing_kv.llm_model_func if hasattr(hashing_kv, "llm_model_func") else None
)
quantized, min_val, max_val = quantize_embedding(current_embedding[0])
best_cached_response = await get_best_cached_response(
hashing_kv,