File size: 3,110 Bytes
86ea792 2644024 3c09fac 2644024 86ea792 2644024 86ea792 133c54b 86ea792 3c09fac 194886f 133c54b 194886f 53e1948 194886f 3c09fac 133c54b 3c09fac 86ea792 53e1948 86ea792 3c09fac 53e1948 3c09fac 194886f 3c09fac 86ea792 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 |
from openai import OpenAI # Assuming Nvidia client is available in the same library, adjust if necessary
import streamlit as st
import os
from datetime import datetime
# Initialize Nvidia client
client = OpenAI(
base_url="https://integrate.api.nvidia.com/v1", # Nvidia API endpoint
api_key=os.environ["NVIDIA_API_KEY"] # Nvidia API Key from Streamlit secrets
)
st.title("ChatGPT-like clone with Nvidia Nemotron 70B Model")
# Sidebar with instructions and Clear Session button
with st.sidebar:
# Instruction
st.markdown("### Instructions 🤖\nThis is a basic chatbot. Ask anything, and the AI will try to help you! The app is supported by Yiqiao Yin.")
# Add a section to ask the user for the response length
st.markdown("#### Select the desired length of the AI response:")
response_length = st.radio(
"How detailed do you want the response to be?",
('Efficient', 'Medium', 'Academic')
)
# Set max_tokens based on user selection
if response_length == 'Efficient':
max_tokens = 200
elif response_length == 'Medium':
max_tokens = 600
else: # 'Academic'
max_tokens = 1024
# Clear
if st.button("Clear Session"):
st.session_state.clear()
st.write(f"Copyright © 2010-{datetime.now().year} Present Yiqiao Yin")
# Initialize session state variables if not already present
if "nvidia_model" not in st.session_state:
st.session_state["nvidia_model"] = "nvidia/llama-3.1-nemotron-70b-instruct"
if "messages" not in st.session_state:
# Adding the initial system message
st.session_state.messages = [{"role": "system", "content": "You are a helpful assistant."}]
# Render the chat history
for message in st.session_state.messages:
with st.chat_message(message["role"]):
st.markdown(message["content"])
# Get new user input
if prompt := st.chat_input("What is up?"):
# Add user message to the session state
st.session_state.messages.append({"role": "user", "content": prompt})
with st.chat_message("user"):
st.markdown(prompt)
# Display assistant's message while waiting for the response
with st.chat_message("assistant"):
with st.spinner("The assistant is thinking... Please wait."):
# Create Nvidia completion request with full conversation history
stream = client.chat.completions.create(
model=st.session_state["nvidia_model"],
messages=st.session_state.messages, # Include all previous messages in the API call
temperature=0.5,
top_p=0.7,
max_tokens=max_tokens,
stream=True,
)
response_chunks = []
for chunk in stream:
if chunk.choices[0].delta.content is not None:
response_chunks.append(chunk.choices[0].delta.content)
response = "".join(response_chunks)
st.markdown(response)
# Store the assistant response in the session state
st.session_state.messages.append({"role": "assistant", "content": response})
|