Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -66,23 +66,28 @@ def create_db(splits):
|
|
66 |
vectordb = FAISS.from_documents(splits, embeddings)
|
67 |
return vectordb
|
68 |
|
69 |
-
def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db):
|
70 |
-
"""Initialize the LLM chain"""
|
|
|
71 |
llm = HuggingFaceEndpoint(
|
72 |
repo_id=llm_model,
|
73 |
huggingfacehub_api_token=api_token,
|
74 |
temperature=temperature,
|
75 |
-
|
76 |
-
top_k
|
77 |
)
|
78 |
|
|
|
79 |
memory = ConversationBufferMemory(
|
80 |
memory_key="chat_history",
|
81 |
output_key='answer',
|
82 |
return_messages=True
|
83 |
)
|
84 |
|
|
|
85 |
retriever = vector_db.as_retriever()
|
|
|
|
|
86 |
qa_chain = ConversationalRetrievalChain.from_llm(
|
87 |
llm,
|
88 |
retriever=retriever,
|
@@ -93,6 +98,7 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db):
|
|
93 |
)
|
94 |
return qa_chain
|
95 |
|
|
|
96 |
def format_chat_history(message, chat_history):
|
97 |
"""Format chat history for the LLM"""
|
98 |
formatted_chat_history = []
|
|
|
66 |
vectordb = FAISS.from_documents(splits, embeddings)
|
67 |
return vectordb
|
68 |
|
69 |
+
def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, api_token):
|
70 |
+
"""Initialize the LLM chain with a HuggingFace model"""
|
71 |
+
# Use valid Hugging Face parameters. `max_length` might be the correct field instead of `max_new_tokens`
|
72 |
llm = HuggingFaceEndpoint(
|
73 |
repo_id=llm_model,
|
74 |
huggingfacehub_api_token=api_token,
|
75 |
temperature=temperature,
|
76 |
+
max_length=max_tokens, # Adjusted from max_new_tokens to max_length
|
77 |
+
# Remove top_k as it may not be valid or handled differently
|
78 |
)
|
79 |
|
80 |
+
# Set up memory for conversation
|
81 |
memory = ConversationBufferMemory(
|
82 |
memory_key="chat_history",
|
83 |
output_key='answer',
|
84 |
return_messages=True
|
85 |
)
|
86 |
|
87 |
+
# Ensure vector_db is used as a retriever
|
88 |
retriever = vector_db.as_retriever()
|
89 |
+
|
90 |
+
# Initialize ConversationalRetrievalChain using LLM and the retriever
|
91 |
qa_chain = ConversationalRetrievalChain.from_llm(
|
92 |
llm,
|
93 |
retriever=retriever,
|
|
|
98 |
)
|
99 |
return qa_chain
|
100 |
|
101 |
+
|
102 |
def format_chat_history(message, chat_history):
|
103 |
"""Format chat history for the LLM"""
|
104 |
formatted_chat_history = []
|