app.py

import gradio as gr
from huggingface_hub import InferenceClient
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel, PeftConfig
impot os
client = InferenceClient(api_key=(os.getenv('HF_TOKEN')))

tokenizer = AutoTokenizer.from_pretrained("BeastGokul/Bio-Mistral-7B-finetuned")
base_model = AutoModelForCausalLM.from_pretrained("BeastGokul/Bio-Mistral-7B-finetuned")
base_model.resize_token_embeddings(len(tokenizer))
model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.3")

def get_model_response(user_query):
    messages = [
        {
            "role": "user",
            "content": user_query
        }
        {
            "role": "biomedical assistant",
            "content": ""
        }
    ]
    
    
    response = client.chat_completions.create(
        model=model,
        messages=messages,
        max_tokens=500
    )
    
    # Extract and return the response content
    return response.choices[0].message['content']

with gr.Blocks() as demo:
    gr.Markdown("# Biomedical Query Chatbot")
    with gr.Row():
        user_input = gr.Textbox(
            placeholder="Enter your biomedical query...",
            label="Your Query",
            lines=1
        )
        submit_button = gr.Button("Submit")
        
    output = gr.Textbox(label="Response from Biomedical Assistant")

    submit_button.click(get_model_response, inputs=user_input, outputs=output)

# Launch the app
demo.launch()