import gradio as gr
from ctransformers import AutoModelForCausalLM
from huggingface_hub import hf_hub_download
model_name = "Hemanth-thunder/Tamil-Mistral-7B-Instruct-v0.1"
model_file = "tamil-mistral-7b-instruct-v0.1.Q4_K_M.gguf" 

model_path = hf_hub_download(model_name, filename=model_file)


 

llm = AutoModelForCausalLM.from_pretrained(model_name, model_file=model_file, 
                                           model_type="mistral", gpu_layers=0)
def alternatingly_agree(message, history):
    outputs = []
    prompt = """<s> சரியான பதிலுடன் வேலையை வெற்றிகரமாக முடிக்க. தேவையான தகவலை உள்ளிடவும்.

    ### Instruction:
    {}

    ### Response:
    """


    prompt = prompt.format(message)
    result = llm(prompt,max_new_tokens=50,temperature=0.7,stream=True)
    for token in result:
        outputs.append(token)
        yield "".join(outputs)
 
 

gr.ChatInterface(alternatingly_agree).launch()