Chat_Flash / app.py
KingNish's picture
Update app.py
42047b7 verified
raw
history blame
975 Bytes
import gradio as gr
from huggingface_hub import InferenceClient
client = InferenceClient("google/gemma-1.1-2b-it")
system_instructions = "[SYSTEM] Your task is to Answer the question. Keep conversation very short, clear and concise. The expectation is that you will avoid introductions and start answering the query directly, Only answer the question asked by user, Do not say unnecessary things.[QUESTION]"
def models(Query):
messages = []
messages.append({"role": "user", "content": Query})
Response = ""
for message in client.chat_completion(
messages,
max_tokens=2048,
stream=True
):
token = message.choices[0].delta.content
Response += token
yield Response
description="# Chat GO\n### Enter your query and Press enter and get response faster than groq"
demo = gr.Interface(description=description,fn=models, inputs=["text"], outputs="text")
demo.queue(max_size=300000)
demo.launch()