File size: 1,840 Bytes
9569dec 21993d9 9569dec 21993d9 9569dec 21993d9 9569dec 21993d9 9569dec 21993d9 9569dec 21993d9 9569dec 21993d9 b2bd387 21993d9 efd03ad |
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 |
import streamlit as st
from huggingface_hub import InferenceClient
# Initialize the client
client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
# Function to format the prompt
def format_prompt(message, history):
prompt = "<s>"
for user_prompt, bot_response in history:
prompt += f"[INST] {user_prompt} [/INST]"
prompt += f" {bot_response}</s> "
prompt += f"[INST] {message} [/INST]"
return prompt
# Function to generate response
def generate(prompt, history, temperature=0.2, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0):
temperature = max(float(temperature), 1e-2)
top_p = float(top_p)
generate_kwargs = dict(
temperature=temperature,
max_new_tokens=max_new_tokens,
top_p=top_p,
repetition_penalty=repetition_penalty,
do_sample=True,
seed=42,
)
formatted_prompt = format_prompt(prompt, history)
stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
output = ""
for response in stream:
output += response.token.text
return output
# Streamlit interface
st.title("Mistral 8x7b Chat")
# Chat history
if 'history' not in st.session_state:
st.session_state.history = []
# User input
user_input = st.text_input("Your message:", key="user_input")
# Generate response and update history
if st.button("Send"):
if user_input:
bot_response = generate(user_input, st.session_state.history)
st.session_state.history.append((user_input, bot_response))
# st.session_state.user_input = ""
# Display conversation
chat_text = ""
for user_msg, bot_msg in st.session_state.history:
chat_text += f"You: {user_msg}\nBot: {bot_msg}\n\n"
st.text_area("Chat", value=chat_text, height=300, disabled=False)
|