| import os |
| import torch |
| from transformers import AutoModelForCausalLM, AutoTokenizer |
| import gradio as gr |
| model_id = "microsoft/DialoGPT-small" |
| model = None |
| tokenizer = None |
| print(f"\nLoading model: {model_id}...") |
| try: |
| tokenizer = AutoTokenizer.from_pretrained(model_id) |
| model = AutoModelForCausalLM.from_pretrained(model_id) |
| model.eval() |
| print("Model loaded successfully!") |
| except Exception as e: |
| print(f"Error loading model: {e}") |
| print("Ensure you are using a CPU-compatible model and that internet access is available.") |
| def medical_qna(question): |
| if model is None or tokenizer is None: |
| return "Error: Model failed to load in a previous step." |
|
|
| try: |
| input_ids = tokenizer.encode(question + tokenizer.eos_token, return_tensors="pt") |
| with torch.no_grad(): |
| output_ids = model.generate( |
| input_ids, |
| max_new_tokens=150, |
| pad_token_id=tokenizer.eos_token_id, |
| do_sample=True, |
| top_k=50, |
| top_p=0.95, |
| temperature=0.7 |
| ) |
|
|
| generated_text = tokenizer.decode(output_ids[:, input_ids.shape[-1]:][0], skip_special_tokens=True) |
| thinking_process = "Analyzing symptoms and context..." |
| final_response = generated_text.strip() |
| formatted_output = f"**Thinking Process:**\n{thinking_process}\n\n**Final Response:**\n{final_response}" |
| return formatted_output |
| except Exception as e: |
| return f"An error occurred during processing: {e}" |
| if model is not None and tokenizer is not None: |
| print("\nSetting up Gradio interface...") |
| try: |
| interface = gr.Interface( |
| fn=medical_qna, |
| inputs=gr.Textbox(lines=3, placeholder="Enter medical question here...", label="Your Medical Question"), |
| outputs=gr.Markdown(label="Chatbot Response"), |
| title="Lightweight Medical Chatbot", |
| description="Ask a medical-style question. This uses a small model for CPU usage. Not for real medical advice." |
| ) |
| print("Launching Gradio interface...") |
| interface.launch() |
|
|
| except Exception as e: |
| print(f"Error launching Gradio interface: {e}") |
| else: |
| print("\nModel was not loaded. Skipping Gradio interface setup.") |
