Spaces:
Sleeping
Sleeping
Update edubot.py
Browse files
edubot.py
CHANGED
@@ -4,7 +4,7 @@ from langchain.vectorstores import FAISS
|
|
4 |
from langchain.llms import CTransformers
|
5 |
from langchain.chains import RetrievalQA
|
6 |
from config import *
|
7 |
-
|
8 |
class EduBotCreator:
|
9 |
|
10 |
def __init__(self):
|
@@ -18,12 +18,12 @@ class EduBotCreator:
|
|
18 |
self.model_type = MODEL_TYPE
|
19 |
self.max_new_tokens = MAX_NEW_TOKENS
|
20 |
self.temperature = TEMPERATURE
|
21 |
-
|
22 |
def create_custom_prompt(self):
|
23 |
custom_prompt_temp = PromptTemplate(template=self.prompt_temp,
|
24 |
input_variables=self.input_variables)
|
25 |
return custom_prompt_temp
|
26 |
-
|
27 |
def load_llm(self):
|
28 |
llm = CTransformers(
|
29 |
model = self.model_ckpt,
|
@@ -32,7 +32,7 @@ class EduBotCreator:
|
|
32 |
temperature = self.temperature
|
33 |
)
|
34 |
return llm
|
35 |
-
|
36 |
def load_vectordb(self):
|
37 |
hfembeddings = HuggingFaceEmbeddings(
|
38 |
model_name=self.embedder,
|
@@ -41,7 +41,8 @@ class EduBotCreator:
|
|
41 |
|
42 |
vector_db = FAISS.load_local(self.vector_db_path, hfembeddings)
|
43 |
return vector_db
|
44 |
-
|
|
|
45 |
def create_bot(self, custom_prompt, vectordb, llm):
|
46 |
retrieval_qa_chain = RetrievalQA.from_chain_type(
|
47 |
llm=llm,
|
|
|
4 |
from langchain.llms import CTransformers
|
5 |
from langchain.chains import RetrievalQA
|
6 |
from config import *
|
7 |
+
import streamlit as st
|
8 |
class EduBotCreator:
|
9 |
|
10 |
def __init__(self):
|
|
|
18 |
self.model_type = MODEL_TYPE
|
19 |
self.max_new_tokens = MAX_NEW_TOKENS
|
20 |
self.temperature = TEMPERATURE
|
21 |
+
|
22 |
def create_custom_prompt(self):
|
23 |
custom_prompt_temp = PromptTemplate(template=self.prompt_temp,
|
24 |
input_variables=self.input_variables)
|
25 |
return custom_prompt_temp
|
26 |
+
@st.cache_resource()
|
27 |
def load_llm(self):
|
28 |
llm = CTransformers(
|
29 |
model = self.model_ckpt,
|
|
|
32 |
temperature = self.temperature
|
33 |
)
|
34 |
return llm
|
35 |
+
@st.cache_resource()
|
36 |
def load_vectordb(self):
|
37 |
hfembeddings = HuggingFaceEmbeddings(
|
38 |
model_name=self.embedder,
|
|
|
41 |
|
42 |
vector_db = FAISS.load_local(self.vector_db_path, hfembeddings)
|
43 |
return vector_db
|
44 |
+
|
45 |
+
|
46 |
def create_bot(self, custom_prompt, vectordb, llm):
|
47 |
retrieval_qa_chain = RetrievalQA.from_chain_type(
|
48 |
llm=llm,
|