Update app.py

app.py

@@ -7,7 +7,7 @@ import chromadb
 from llama_index.vector_stores.chroma import ChromaVectorStore
 from llama_index.llms.ollama import Ollama
 
-#from llama_index.llms.huggingface import HuggingFaceLLM
+from llama_index.llms.huggingface import HuggingFaceLLM
 
 from llama_index.core import Settings
 
@@ -32,7 +32,18 @@ pipeline = IngestionPipeline(
 index = VectorStoreIndex.from_vector_store(vector_store, embed_model=embed_model)
 
 # Utwórz silnik zapytań
-llm = Ollama(model="qwen2:7b")
+# huggingface
+from transformers import AutoTokenizer
+
+# Settings.tokenizer = AutoTokenizer.from_pretrained(
+#     "Qwen/Qwen2-7B-Instruct"
+# )
+
+# Load the correct tokenizer for Qwen/Qwen2-7B-Instruct
+tokeni = AutoTokenizer.from_pretrained("Qwen/Qwen2-0.5B")
+
+llm = HuggingFaceLLM(model_name="Qwen/Qwen2-0.5B", tokenizer=tokeni)
+# print(llm._tokenizer)
 query_engine = index.as_query_engine(
     llm=llm,
     response_mode='compact')
@@ -67,6 +78,4 @@ if st.session_state.messages[-1]["role"] != "assistant":
             st.write(content)  # Wyświetl całą treść w Streamlit
 
     message = {"role": "assistant", "content": content}  # Zapisz całą treść w wiadomości
-    st.session_state.messages.append(message)
-
-
+    st.session_state.messages.append(message)