app1.py

from ctransformers import AutoModelForCausalLM
import gradio as gr

greety = """
 A special thanks to the CCL AI team members who made a significant contribution to this project.
"""

llm = AutoModelForCausalLM.from_pretrained("pt_merge_model_v3.Q4_K_M.gguf",
model_type='llama',
max_new_tokens = 512,
threads = 3,
)

def stream(prompt, UL):
    system_prompt = 'You are a helpful AI assistant'
    prompt =  f"<s>[INST]<<SYS>>\n + {system_prompt} + <</SYS>>\n{prompt.strip()}[/INST]"
    return llm(prompt)

css = """
  h1 {
  text-align: center;
}
#duplicate-button {
  margin: auto;
  color: white;
  background: #1565c0;
  border-radius: 100vh;
}
.contain {
  max-width: 900px;
  margin: auto;
  padding-top: 1.5rem;
}
"""

chat_interface = gr.ChatInterface(
    fn=stream,
    #additional_inputs_accordion_name = "Credentials",
    #additional_inputs=[
    #     gr.Textbox(label="OpenAI Key", lines=1),
    #     gr.Textbox(label="Linkedin Access Token", lines=1),
    #],
    stop_btn=None,
    examples=[
        ["explain Large language model"],
        ["what is quantum computing"],
        ["Explain what is atomic reaction"]
    ],
)

with gr.Blocks(css=css) as demo:
    gr.HTML("<h1><center>RoyalGPT Free LLM Deployment Space<h1><center>")
    gr.HTML("<h3><center><a href='https://royalgpt.xyz/'>RoyalGPT</a>💬<h3><center>")
    gr.DuplicateButton(value="Duplicate Space for private use", elem_id="duplicate-button")
    chat_interface.render()
    gr.Markdown(greety)
    
if __name__ == "__main__":
    demo.queue(max_size=10).launch(share=True)