pengjunjie/ragflow
1
0
Fork 1
mirror of https://github.com/infiniflow/ragflow.git synced 2025-12-08 20:42:30 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix: The max tokens defined by the tenant are not used (#4297) (#2817) (#5066)

Browse Source 
### What problem does this PR solve?

Fix: The max tokens defined by the tenant are not used (#4297) (#2817)


### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)

---------

Co-authored-by: Kevin Hu <kevinhu.sh@gmail.com>
This commit is contained in:
zhxlp
2025-02-18 13:42:22 +08:00
committed by GitHub
parent 3e0bc9e36b
commit 00c7ddbc9b
2 changed files with 15 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
api/db/services/llm_service.py
View File

@ -86,8 +86,7 @@ class TenantLLMService(CommonService):
@classmethod
@DB.connection_context()
def model_instance(cls, tenant_id, llm_type,
llm_name=None, lang="Chinese"):
def get_model_config(cls, tenant_id, llm_type, llm_name=None):
e, tenant = TenantService.get_by_id(tenant_id)
if not e:
raise LookupError("Tenant not found")
@ -124,7 +123,13 @@ class TenantLLMService(CommonService):
if not mdlnm:
raise LookupError(f"Type of {llm_type} model is not set.")
raise LookupError("Model({}) not authorized".format(mdlnm))
return model_config
@classmethod
@DB.connection_context()
def model_instance(cls, tenant_id, llm_type,
llm_name=None, lang="Chinese"):
model_config = TenantLLMService.get_model_config(tenant_id, llm_type, llm_name)
if llm_type == LLMType.EMBEDDING.value:
if model_config["llm_factory"] not in EmbeddingModel:
return
@ -228,10 +233,8 @@ class LLMBundle(object):
tenant_id, llm_type, llm_name, lang=lang)
assert self.mdl, "Can't find model for {}/{}/{}".format(
tenant_id, llm_type, llm_name)
self.max_length = 8192
for lm in LLMService.query(llm_name=llm_name):
self.max_length = lm.max_tokens
break
model_config = TenantLLMService.get_model_config(tenant_id, llm_type, llm_name)
self.max_length = model_config.get("max_tokens", 8192)
def encode(self, texts: list):
embeddings, used_tokens = self.mdl.encode(texts)