Fix llm_model_func retrieval error.
This commit is contained in:
@@ -368,7 +368,7 @@ async def get_best_cached_response(
|
|||||||
original_prompt=None,
|
original_prompt=None,
|
||||||
cache_type=None,
|
cache_type=None,
|
||||||
) -> Union[str, None]:
|
) -> Union[str, None]:
|
||||||
logger.debug(f"get_best_cached_response: mode={mode} cache_type={cache_type}")
|
logger.debug(f"get_best_cached_response: mode={mode} cache_type={cache_type} use_llm_check={use_llm_check}")
|
||||||
mode_cache = await hashing_kv.get_by_id(mode)
|
mode_cache = await hashing_kv.get_by_id(mode)
|
||||||
if not mode_cache:
|
if not mode_cache:
|
||||||
return None
|
return None
|
||||||
@@ -511,11 +511,7 @@ async def handle_cache(
|
|||||||
if is_embedding_cache_enabled:
|
if is_embedding_cache_enabled:
|
||||||
# Use embedding cache
|
# Use embedding cache
|
||||||
current_embedding = await hashing_kv.embedding_func([prompt])
|
current_embedding = await hashing_kv.embedding_func([prompt])
|
||||||
llm_model_func = (
|
llm_model_func = hashing_kv.global_config.get('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])
|
quantized, min_val, max_val = quantize_embedding(current_embedding[0])
|
||||||
best_cached_response = await get_best_cached_response(
|
best_cached_response = await get_best_cached_response(
|
||||||
hashing_kv,
|
hashing_kv,
|
||||||
|
Reference in New Issue
Block a user