refactor: Update llm.py and prompt.py files
Browse files- libs/llm.py +8 -3
- libs/prompt.py +1 -1
libs/llm.py
CHANGED
@@ -10,7 +10,7 @@ from langchain.callbacks.base import BaseCallbackHandler
|
|
10 |
|
11 |
class StreamCallback(BaseCallbackHandler):
|
12 |
def on_llm_new_token(self, token: str, **kwargs):
|
13 |
-
print(token, end="", flush=True)
|
14 |
|
15 |
|
16 |
def get_llm(streaming=True):
|
@@ -46,8 +46,13 @@ def get_llm(streaming=True):
|
|
46 |
streaming=streaming,
|
47 |
callbacks=[StreamCallback()],
|
48 |
),
|
49 |
-
|
50 |
-
model="EEVE-Korean-10.8B
|
|
|
|
|
|
|
|
|
|
|
51 |
streaming=streaming,
|
52 |
callbacks=[StreamCallback()],
|
53 |
),
|
|
|
10 |
|
11 |
class StreamCallback(BaseCallbackHandler):
|
12 |
def on_llm_new_token(self, token: str, **kwargs):
|
13 |
+
# print(token, end="", flush=True)
|
14 |
|
15 |
|
16 |
def get_llm(streaming=True):
|
|
|
46 |
streaming=streaming,
|
47 |
callbacks=[StreamCallback()],
|
48 |
),
|
49 |
+
eeve=ChatOllama(
|
50 |
+
model="EEVE-Korean-10.8B",
|
51 |
+
streaming=streaming,
|
52 |
+
callbacks=[StreamCallback()],
|
53 |
+
),
|
54 |
+
gemma2=ChatOllama(
|
55 |
+
model="gemma2",
|
56 |
streaming=streaming,
|
57 |
callbacks=[StreamCallback()],
|
58 |
),
|
libs/prompt.py
CHANGED
@@ -28,4 +28,4 @@ def get_prompt(chat_history):
|
|
28 |
MessagesPlaceholder(variable_name="history"),
|
29 |
("human", "{question}"),
|
30 |
]
|
31 |
-
).partial(history=chat_history
|
|
|
28 |
MessagesPlaceholder(variable_name="history"),
|
29 |
("human", "{question}"),
|
30 |
]
|
31 |
+
).partial(history=chat_history)
|