Delete test.py
Browse files
test.py
DELETED
@@ -1,37 +0,0 @@
|
|
1 |
-
import langchain as lc
|
2 |
-
from langchain import PromptTemplate, LLMChain
|
3 |
-
from langchain_community.llms import LlamaCpp
|
4 |
-
from langchain.chains import ConversationChain
|
5 |
-
import gradio as gr
|
6 |
-
|
7 |
-
# Define the prompt template
|
8 |
-
template = """
|
9 |
-
<s>사용자 질문에 알맞은 답변을 해주세요.
|
10 |
-
|
11 |
-
{history}
|
12 |
-
<|im_start|>user
|
13 |
-
{input}
|
14 |
-
<|im_end|>
|
15 |
-
<|im_start|>assistant
|
16 |
-
"""
|
17 |
-
|
18 |
-
# Make sure the model path is correct for your system!
|
19 |
-
llm = LlamaCpp(
|
20 |
-
model_path="models/10.7B/EEVE_ggml-model-Q5_K_M.gguf",
|
21 |
-
lora_path="loras/240417_275dw_2ep_new_prompt/ggml-adapter-model.bin",
|
22 |
-
temperature=0.1,
|
23 |
-
max_tokens=4096,
|
24 |
-
early_stopping=True,
|
25 |
-
do_sample=True,
|
26 |
-
repetition_penalty=1.17,
|
27 |
-
top_k=40,
|
28 |
-
top_p=0.1,
|
29 |
-
stop=['<|im_end|>']
|
30 |
-
)
|
31 |
-
|
32 |
-
def predict(input, history):
|
33 |
-
prompt = PromptTemplate(input_variables=["input","history"], template=template)
|
34 |
-
chain = LLMChain(prompt=prompt, llm=llm)
|
35 |
-
return chain.invoke(input=input, history=history)
|
36 |
-
|
37 |
-
gr.ChatInterface(predict).queue().launch(share=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|