llama / app.py
Nexchan's picture
Update app.py
bf73f38 verified
raw
history blame
2.77 kB
import gradio as gr
from huggingface_hub import InferenceClient
client = InferenceClient("meta-llama/Meta-Llama-3-8B-Instruct")
def chat_llama(chat_history):
chat_completion = client.chat_completion(
messages=chat_history,
max_tokens=500,
# stream=True
)
# Append the assistant's response to chat history
chat_history.append({"role": "assistant", "content": chat_completion.choices[0].message.content})
# Format chat_history as a list of tuples
formatted_history = [(chat_history[i*2]["content"], chat_history[i*2+1]["content"])
for i in range(len(chat_history)//2)]
return formatted_history
def chat_mem(message, chat_history):
chat_history_role = [{"role": "system", "content": "You are a helpful assistant."}]
if chat_history:
for user_message, assistant_message in chat_history:
chat_history_role.append({"role": "user", "content": user_message})
chat_history_role.append({"role": "assistant", "content": assistant_message})
chat_history_role.append({"role": "user", "content": message})
chat_completion = client.chat_completion(
messages=chat_history_role,
max_tokens=500,
# stream=True
)
assistant_message = chat_completion.choices[0].message.content
chat_history_role.append({"role": "assistant", "content": assistant_message})
# Format chat_history as a list of tuples
formatted_history = [(chat_history_role[i*2]["content"], chat_history_role[i*2+1]["content"])
for i in range(len(chat_history_role)//2)]
return "", formatted_history
with gr.Blocks() as demo:
with gr.Row():
with gr.Column():
chatbot = gr.Chatbot()
msg = gr.Textbox(placeholder="Type a message...", interactive=True)
with gr.Row():
clear = gr.ClearButton([msg, chatbot], icon="https://img.icons8.com/?size=100&id=Xnx8cxDef16O&format=png&color=000000")
send_btn = gr.Button("Send", variant='primary', icon="https://img.icons8.com/?size=100&id=g8ltXTwIfJ1n&format=png&color=000000")
msg.submit(fn=chat_mem, inputs=[msg, chatbot], outputs=[msg, chatbot])
send_btn.click(fn=chat_mem, inputs=[msg, chatbot], outputs=[msg, chatbot])
with gr.Column():
gr.Markdown("### API Testing")
json_input = gr.Textbox(label="Input JSON", placeholder='Enter JSON here', lines=10)
json_output = gr.Textbox(label="Output JSON", lines=10, interactive=False)
test_btn = gr.Button("Test API")
test_btn.click(fn=chat_llama, inputs=json_input, outputs=json_output)
if __name__ == "__main__":
demo.launch()