Update main.py

main.py

@@ -17,13 +17,13 @@ from transformers import AutoModelForSequenceClassification, AutoTokenizer
 
 
 #
-# tokenizer = AutoTokenizer.from_pretrained("google/flan-t5-base")
-#
-# model = AutoModelForSeq2SeqLM.from_pretrained("google/flan-t5-base")
+tokenizer = AutoTokenizer.from_pretrained("google/flan-t5-base")
+
+model = AutoModelForSeq2SeqLM.from_pretrained("google/flan-t5-base")
 
-model_id = "lamdao/lora-trained-xl-colab"
-tokenizer = AutoTokenizer.from_pretrained(model_id)
-model = AutoModelForCausalLM.from_pretrained(model_id)
+# model_id = "lamdao/lora-trained-xl-colab"
+# tokenizer = AutoTokenizer.from_pretrained(model_id)
+# model = AutoModelForCausalLM.from_pretrained(model_id)
 
 pipeline = pipeline(
     "text-generation",
@@ -48,14 +48,14 @@ local_llm = HuggingFacePipeline(pipeline=pipeline)
 loader = PyPDFLoader('bipolar.pdf')
 # loader = TextLoader('info.txt')
 document = loader.load()
-text_spliter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
+text_spliter = CharacterTextSplitter(chunk_size=100, chunk_overlap=0)
 texts = text_spliter.split_documents(document)
 embedding = HuggingFaceInstructEmbeddings()
 docsearch = Chroma.from_documents(texts, embedding, persist_directory='db')
 
 retriever = docsearch.as_retriever(search_kwargs={"k": 3})
 qa_chain = RetrievalQA.from_chain_type(llm=local_llm,
-                                       chain_type="map_reduce",
+                                       chain_type="stuff",
                                        retriever=retriever,
                                        return_source_documents=True)
 # question = input('prompt: ')
@@ -66,7 +66,7 @@ def gradinterface(query,history):
     return result['result']
 
 
-demo = gr.ChatInterface(fn=gradinterface, title='OUR_OWN_BOT')
+demo = gr.ChatInterface(fn=gradinterface, title='OUR_BOT')
 
 if __name__ == "__main__":
     demo.launch(share=True)