AI/ragflow
1
0
Fork 0
mirror of https://git.mirrors.martin98.com/https://github.com/infiniflow/ragflow.git synced 2025-08-02 04:30:38 +08:00
Code Issues Packages Projects Releases Wiki Activity

Feat: add token comsumption & speed to little lamp. (#6077)

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

#6059

### Type of change

- [x] New Feature (non-breaking change which adds functionality)
This commit is contained in:
Kevin Hu 2025-03-14 13:37:31 +08:00 committed by GitHub
parent c85b468b8d
commit 42eb99554f
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 20 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
agent/component/base.py
View File

@ -521,7 +521,7 @@ class ComponentBase(ABC):
if u.lower().find("answer") >= 0: if u.lower().find("answer") >= 0:
for r, c in self._canvas.history[::-1]: for r, c in self._canvas.history[::-1]:
if r == "user": if r == "user":
upstream_outs.append(pd.DataFrame([{"content": c, "component_id": u}])) upstream_outs.append(pd.DataFrame([{"content": f"USER:{c}", "component_id": u}]))
break break
break break
if self.component_name.lower().find("answer") >= 0 and self.get_component_name(u) in ["relevant"]: if self.component_name.lower().find("answer") >= 0 and self.get_component_name(u) in ["relevant"]:

19
api/db/services/dialog_service.py
View File

@ -304,8 +304,25 @@ def chat(dialog, messages, stream=True, **kwargs):
retrieval_time_cost = (retrieval_ts - generate_keyword_ts) * 1000 retrieval_time_cost = (retrieval_ts - generate_keyword_ts) * 1000
generate_result_time_cost = (finish_chat_ts - retrieval_ts) * 1000 generate_result_time_cost = (finish_chat_ts - retrieval_ts) * 1000
tk_num = num_tokens_from_string(think+answer)
prompt += "\n\n### Query:\n%s" % " ".join(questions) prompt += "\n\n### Query:\n%s" % " ".join(questions)
prompt = f"{prompt}\n\n - Total: {total_time_cost:.1f}ms\n - Check LLM: {check_llm_time_cost:.1f}ms\n - Create retriever: {create_retriever_time_cost:.1f}ms\n - Bind embedding: {bind_embedding_time_cost:.1f}ms\n - Bind LLM: {bind_llm_time_cost:.1f}ms\n - Tune question: {refine_question_time_cost:.1f}ms\n - Bind reranker: {bind_reranker_time_cost:.1f}ms\n - Generate keyword: {generate_keyword_time_cost:.1f}ms\n - Retrieval: {retrieval_time_cost:.1f}ms\n - Generate answer: {generate_result_time_cost:.1f}ms" prompt = (
f"{prompt}\n\n"
"## Time elapsed:\n"
f" - Total: {total_time_cost:.1f}ms\n"
f" - Check LLM: {check_llm_time_cost:.1f}ms\n"
f" - Create retriever: {create_retriever_time_cost:.1f}ms\n"
f" - Bind embedding: {bind_embedding_time_cost:.1f}ms\n"
f" - Bind LLM: {bind_llm_time_cost:.1f}ms\n"
f" - Tune question: {refine_question_time_cost:.1f}ms\n"
f" - Bind reranker: {bind_reranker_time_cost:.1f}ms\n"
f" - Generate keyword: {generate_keyword_time_cost:.1f}ms\n"
f" - Retrieval: {retrieval_time_cost:.1f}ms\n"
f" - Generate answer: {generate_result_time_cost:.1f}ms\n\n"
"## Token usage:\n"
f" - Generated tokens(approximately): {tk_num}\n"
f" - Token speed: {int(tk_num/(generate_result_time_cost/1000.))}/s"
)
return {"answer": think+answer, "reference": refs, "prompt": re.sub(r"\n", " \n", prompt), "created_at": time.time()} return {"answer": think+answer, "reference": refs, "prompt": re.sub(r"\n", " \n", prompt), "created_at": time.time()}
if stream: if stream: