app.py

# --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 
# 
# This space is created by SANJOG GHONGE for testing and learning purpose.
#
# If you want to remove this space or credits please contact me on my email id [[email protected]].
#
# Citation : @misc{qvq-72b-preview,
#               title = {QVQ: To See the World with Wisdom},
#               url = {https://qwenlm.github.io/blog/qvq-72b-preview/},
#               author = {Qwen Team},
#               month = {December},
#               year = {2024}
#                  }

#           @article{Qwen2VL,
#               title={Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution},
#               author={Wang, Peng and Bai, Shuai and Tan, Sinan and Wang, Shijie and Fan, Zhihao and Bai, 
#               Jinze and Chen, Keqin and Liu, Xuejing and Wang, Jialin and Ge, Wenbin and Fan, Yang and Dang, 
#               Kai and Du, Mengfei and Ren, Xuancheng and Men, Rui and Liu, Dayiheng and Zhou, Chang and Zhou, 
#               Jingren and Lin, Junyang},
#               journal={arXiv preprint arXiv:2409.12191},
#               year={2024}
#                   }
#
# -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------

from transformers import Qwen2VLForConditionalGeneration, AutoProcessor
from qwen_vl_utils import process_vision_info
import gradio as gr
from PIL import Image

# Load the model and processor
model = Qwen2VLForConditionalGeneration.from_pretrained(
    "Qwen/QVQ-72B-Preview", torch_dtype="auto", device_map="auto"
)
processor = AutoProcessor.from_pretrained("Qwen/QVQ-72B-Preview")

# Function to process the image and question
def process_image_and_question(image, question):
    if image is None or question.strip() == "":
        return "Please provide both an image and a question."

    # Prepare the input message
    messages = [
        {
            "role": "system",
            "content": [
                {"type": "text", "text": "You are a helpful and harmless assistant. You are Qwen developed by Alibaba. You should think step-by-step."}
            ],
        },
        {
            "role": "user",
            "content": [
                {"type": "image", "image": image},
                {"type": "text", "text": question},
            ],
        }
    ]

    # Process the inputs
    text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
    image_inputs, video_inputs = process_vision_info(messages)

    inputs = processor(
        text=[text],
        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=8192)
    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] if output_text else "No output generated."

# Define the Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("# Sanjog Image and Question Answering\nProvide an image (JPG/PNG) and a related question to get an answer.")

    with gr.Row():
        with gr.Column():
            image_input = gr.Image(type="pil", label="Upload Image (JPG/PNG)")
            question_input = gr.Textbox(label="Enter your question")

        with gr.Column():
            output_box = gr.Textbox(label="Result", interactive=False)

    with gr.Row():
        clear_button = gr.Button("Clear")
        submit_button = gr.Button("Submit")

    # Define button functionality
    clear_button.click(lambda: (None, "", ""), inputs=[], outputs=[image_input, question_input, output_box])
    submit_button.click(process_image_and_question, inputs=[image_input, question_input], outputs=output_box)

# Launch the interface
demo.launch()


# ------------------------------------------------------------------------------------------------------------------------------------



# import gradio as gr
# from transformers import AutoProcessor, AutoModelForImageTextToText

# # Load the processor and model
# model_name = "Qwen/QVQ-72B-Preview"
# processor = AutoProcessor.from_pretrained(model_name)
# model = AutoModelForImageTextToText.from_pretrained(model_name)

# # Define the prediction function
# def process_image_and_question(image, question):
#     if image is None or not question:
#         return "Please provide both an image and a question."
    
#     # Process the inputs
#     inputs = processor(images=image, text=question, return_tensors="pt")
    
#     # Generate the output
#     outputs = model.generate(**inputs)
#     answer = processor.batch_decode(outputs, skip_special_tokens=True)[0]
    
#     return answer

# # Define the Gradio interface
# with gr.Blocks() as demo:
#     gr.Markdown("# Image and Question Answering\nProvide an image (JPG/PNG) and a related question to get an answer.")

#     with gr.Row():
#         with gr.Column():
#             image_input = gr.Image(type="pil", label="Upload Image (JPG/PNG)")
#             question_input = gr.Textbox(label="Enter your question")

#         with gr.Column():
#             output_box = gr.Textbox(label="Result", interactive=False)

#     with gr.Row():
#         clear_button = gr.Button("Clear")
#         submit_button = gr.Button("Submit")

#     # Define button functionality
#     clear_button.click(lambda: (None, "", ""), inputs=[], outputs=[image_input, question_input, output_box])
#     submit_button.click(process_image_and_question, inputs=[image_input, question_input], outputs=output_box)

# # Launch the interface
# demo.launch()