app.py

import gradio as gr
from transformers import AutoProcessor, LlavaForConditionalGeneration
from PIL import Image
import torch

model_id = "llava-hf/llava-1.5-7b-hf"
processor = AutoProcessor.from_pretrained(model_id)
model = LlavaForConditionalGeneration.from_pretrained(
    model_id,
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
    low_cpu_mem_usage=True,
    device_map="auto",
)

def chat_with_llava(image, question, history=[]):
    if image is None or not question.strip():
        history.append([question, "Please provide both an image and a question."])
        return history

    # Format multimodal prompt
    conversation = [
        {"role": "user", "content": [
            {"type": "text", "text": question},
            {"type": "image"}
        ]}
    ]
    prompt = processor.apply_chat_template(conversation, add_generation_prompt=True)

    # Encode inputs
    inputs = processor(images=image, text=prompt, return_tensors="pt").to(model.device)
    outputs = model.generate(**inputs, max_new_tokens=512)
    answer = processor.decode(outputs[0], skip_special_tokens=True)

    history.append([question, answer])
    return history

chat_interface = gr.ChatInterface(
    fn=chat_with_llava,
    inputs=[gr.Image(type="pil", label="Palm Image"), gr.Textbox(label="Your Question")],
    title="🖐️ AI Palm Reader",
    description="Upload your palm image and ask a question—LLaVA will respond with a palmistry-style reading."
)
chat_interface.launch()