my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	---
#2	title: '🚀 Streamlit'
#3	description: 'Integrate with Streamlit to plug and play with any LLM'
#4	---
#5
#6	In this example, we will learn how to use `mistralai/Mixtral-8x7B-Instruct-v0.1` and Embedchain together with Streamlit to build a simple RAG chatbot.
#7
#8	![Streamlit + Embedchain Demo](https://github.com/embedchain/embedchain/assets/73601258/052f7378-797c-41cf-ac81-f004d0d44dd1)
#9
#10	## Setup
#11
#12	Install Embedchain and Streamlit.
#13	```bash
#14	pip install embedchain streamlit
#15	```
#16	<Tabs>
#17	<Tab title="app.py">
#18	```python
#19	import os
#20	from embedchain import App
#21	import streamlit as st
#22
#23	with st.sidebar:
#24	huggingface_access_token = st.text_input("Hugging face Token", key="chatbot_api_key", type="password")
#25	"[Get Hugging Face Access Token](https://huggingface.co/settings/tokens)"
#26	"[View the source code](https://github.com/embedchain/examples/mistral-streamlit)"
#27
#28
#29	st.title("💬 Chatbot")
#30	st.caption("🚀 An Embedchain app powered by Mistral!")
#31	if "messages" not in st.session_state:
#32	st.session_state.messages = [
#33	{
#34	"role": "assistant",
#35	"content": """
#36	Hi! I'm a chatbot. I can answer questions and learn new things!\n
#37	Ask me anything and if you want me to learn something do `/add <source>`.\n
#38	I can learn mostly everything. :)
#39	""",
#40	}
#41	]
#42
#43	for message in st.session_state.messages:
#44	with st.chat_message(message["role"]):
#45	st.markdown(message["content"])
#46
#47	if prompt := st.chat_input("Ask me anything!"):
#48	if not st.session_state.chatbot_api_key:
#49	st.error("Please enter your Hugging Face Access Token")
#50	st.stop()
#51
#52	os.environ["HUGGINGFACE_ACCESS_TOKEN"] = st.session_state.chatbot_api_key
#53	app = App.from_config(config_path="config.yaml")
#54
#55	if prompt.startswith("/add"):
#56	with st.chat_message("user"):
#57	st.markdown(prompt)
#58	st.session_state.messages.append({"role": "user", "content": prompt})
#59	prompt = prompt.replace("/add", "").strip()
#60	with st.chat_message("assistant"):
#61	message_placeholder = st.empty()
#62	message_placeholder.markdown("Adding to knowledge base...")
#63	app.add(prompt)
#64	message_placeholder.markdown(f"Added {prompt} to knowledge base!")
#65	st.session_state.messages.append({"role": "assistant", "content": f"Added {prompt} to knowledge base!"})
#66	st.stop()
#67
#68	with st.chat_message("user"):
#69	st.markdown(prompt)
#70	st.session_state.messages.append({"role": "user", "content": prompt})
#71
#72	with st.chat_message("assistant"):
#73	msg_placeholder = st.empty()
#74	msg_placeholder.markdown("Thinking...")
#75	full_response = ""
#76
#77	for response in app.chat(prompt):
#78	msg_placeholder.empty()
#79	full_response += response
#80
#81	msg_placeholder.markdown(full_response)
#82	st.session_state.messages.append({"role": "assistant", "content": full_response})
#83	```
#84	</Tab>
#85	<Tab title="config.yaml">
#86	```yaml
#87	app:
#88	config:
#89	name: 'mistral-streamlit-app'
#90
#91	llm:
#92	provider: huggingface
#93	config:
#94	model: 'mistralai/Mixtral-8x7B-Instruct-v0.1'
#95	temperature: 0.1
#96	max_tokens: 250
#97	top_p: 0.1
#98	stream: true
#99
#100	embedder:
#101	provider: huggingface
#102	config:
#103	model: 'sentence-transformers/all-mpnet-base-v2'
#104	```
#105	</Tab>
#106	</Tabs>
#107
#108	## To run it locally,
#109
#110	```bash
#111	streamlit run app.py
#112	```
#113

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public