app.py

import gradio as gr
from transformers import Qwen2_5_VLForConditionalGeneration, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch

# Specify the local cache path for models
local_path = "Qwen/Qwen2.5-VL-7B-Instruct"

# Load model and processor
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
    local_path, torch_dtype="auto", device_map="auto"
)

processor = AutoProcessor.from_pretrained(local_path)

# Function to process image and text and generate the output
@torch.inference_mode()
@spaces.GPU(duration=120)  # Specify a duration to avoid timeout
def generate_output(image, text, button_click):
    # Prepare input data
    messages = [
        {
            "role": "user",
            "content": [
                {"type": "image", "image": image},
                {"type": "text", "text": text},
            ],
        }
    ]
    
    # Prepare inputs for the model
    text_input = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
    image_inputs, video_inputs = process_vision_info(messages)
    inputs = processor(
        text=[text_input],
        images=image_inputs,
        videos=video_inputs,
        padding=True,
        return_tensors="pt",
    )
    inputs = inputs.to("cuda")

    # Generate the output
    generated_ids = model.generate(**inputs, max_new_tokens=128)
    generated_ids_trimmed = [
        out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
    ]
    output_text = processor.batch_decode(
        generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
    )
    return output_text[0]

# Create Gradio interface
iface = gr.Interface(
    fn=generate_output,
    inputs=[
        gr.Image(type="pil", label="Upload Image"),
        gr.Textbox(lines=2, placeholder="Enter a question related to the image", label="Input Text"),
        
    ],
    outputs=gr.Textbox(label="Model Output"),
)

# Launch the Gradio interface
iface.launch()