Fix bug over chunks classification by document in the promp (#4156)
The refactor in 0.15 contains a small bug that eliminate the classification ### What problem does this PR solve ### Type of change - [X] Bug Fix (non-breaking change which fixes an issue)
This commit is contained in:
parent
f13f503952
commit
96da618b6a
1 changed files with 1 additions and 1 deletions
|
|
@ -124,7 +124,7 @@ def kb_prompt(kbinfos, max_tokens):
|
|||
for i, ck in enumerate(kbinfos["chunks"]):
|
||||
if i >= chunks_num:
|
||||
break
|
||||
doc2chunks["docnm_kwd"].append(ck["content_with_weight"])
|
||||
doc2chunks[ck["docnm_kwd"]].append(ck["content_with_weight"])
|
||||
|
||||
knowledges = []
|
||||
for nm, chunks in doc2chunks.items():
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue