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

Pass top_p to ollama (#3744)

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

Pass top_p to ollama. Close #1769

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
This commit is contained in:
Zhichang Yu
2024-11-29 14:52:27 +08:00
committed by GitHub
parent 0a62dd7a7e
commit d94386e00a
1 changed files with 5 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
rag/llm/chat_model.py
View File

@ -356,7 +356,7 @@ class OllamaChat(Base):
options = {} options = {}
if "temperature" in gen_conf: options["temperature"] = gen_conf["temperature"] if "temperature" in gen_conf: options["temperature"] = gen_conf["temperature"]
if "max_tokens" in gen_conf: options["num_predict"] = gen_conf["max_tokens"] if "max_tokens" in gen_conf: options["num_predict"] = gen_conf["max_tokens"]
if "top_p" in gen_conf: options["top_k"] = gen_conf["top_p"] if "top_p" in gen_conf: options["top_p"] = gen_conf["top_p"]
if "presence_penalty" in gen_conf: options["presence_penalty"] = gen_conf["presence_penalty"] if "presence_penalty" in gen_conf: options["presence_penalty"] = gen_conf["presence_penalty"]
if "frequency_penalty" in gen_conf: options["frequency_penalty"] = gen_conf["frequency_penalty"] if "frequency_penalty" in gen_conf: options["frequency_penalty"] = gen_conf["frequency_penalty"]
response = self.client.chat( response = self.client.chat(
@ -376,7 +376,7 @@ class OllamaChat(Base):
options = {} options = {}
if "temperature" in gen_conf: options["temperature"] = gen_conf["temperature"] if "temperature" in gen_conf: options["temperature"] = gen_conf["temperature"]
if "max_tokens" in gen_conf: options["num_predict"] = gen_conf["max_tokens"] if "max_tokens" in gen_conf: options["num_predict"] = gen_conf["max_tokens"]
if "top_p" in gen_conf: options["top_k"] = gen_conf["top_p"] if "top_p" in gen_conf: options["top_p"] = gen_conf["top_p"]
if "presence_penalty" in gen_conf: options["presence_penalty"] = gen_conf["presence_penalty"] if "presence_penalty" in gen_conf: options["presence_penalty"] = gen_conf["presence_penalty"]
if "frequency_penalty" in gen_conf: options["frequency_penalty"] = gen_conf["frequency_penalty"] if "frequency_penalty" in gen_conf: options["frequency_penalty"] = gen_conf["frequency_penalty"]
ans = "" ans = ""
@ -430,7 +430,7 @@ class LocalLLM(Base):
try: try:
self._connection.send(pickle.dumps((name, args, kwargs))) self._connection.send(pickle.dumps((name, args, kwargs)))
return pickle.loads(self._connection.recv()) return pickle.loads(self._connection.recv())
except Exception as e: except Exception:
self.__conn() self.__conn()
raise Exception("RPC connection lost!") raise Exception("RPC connection lost!")
@ -442,7 +442,7 @@ class LocalLLM(Base):
self.client = Client(port=12345, protocol="grpc", asyncio=True) self.client = Client(port=12345, protocol="grpc", asyncio=True)
def _prepare_prompt(self, system, history, gen_conf): def _prepare_prompt(self, system, history, gen_conf):
from rag.svr.jina_server import Prompt, Generation from rag.svr.jina_server import Prompt
if system: if system:
history.insert(0, {"role": "system", "content": system}) history.insert(0, {"role": "system", "content": system})
if "max_tokens" in gen_conf: if "max_tokens" in gen_conf:
@ -450,7 +450,7 @@ class LocalLLM(Base):
return Prompt(message=history, gen_conf=gen_conf) return Prompt(message=history, gen_conf=gen_conf)
def _stream_response(self, endpoint, prompt): def _stream_response(self, endpoint, prompt):
from rag.svr.jina_server import Prompt, Generation from rag.svr.jina_server import Generation
answer = "" answer = ""
try: try:
res = self.client.stream_doc( res = self.client.stream_doc(