app.py

from PIL import Image
import spaces

import gradio as gr

MODEL_ID = "Qwen/Qwen2-VL-7B-Instruct"
MODEL_FINETUNE_ID = "davidr99/qwen2.5-7b-instruct-blackjack"

EXAMPLES = [
    "examples/black_jack_screenshot_1737088587.png",
    "examples/black_jack_screenshot_1737088629.png",
    "examples/black_jack_screenshot_1737088648.png",
    "examples/Screenshot 2024-12-06 220410.png"
]


from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info

model = Qwen2VLForConditionalGeneration.from_pretrained(MODEL_ID, torch_dtype="auto").to('cuda')
model.load_adapter(MODEL_FINETUNE_ID)
processor = AutoProcessor.from_pretrained(MODEL_FINETUNE_ID)

@spaces.GPU(duration=30)
def blackjack_ai(image, question):

    instruction = question
    
    messages = [
        {"role": "system",
            "content": [
            {"type":"text", "text": "You are a blackjack player.  Extract the image into json information."} ]
        },
        {"role": "user", "content": [
            {"type": "image", "image": image},
            {"type": "text", "text": instruction}
        ]}
    ]

    print(messages)

    # Preparation for inference
    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")

    # Inference: Generation of 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

with gr.Blocks() as demo:
   
    image = gr.Image(type="filepath")
    question = gr.Textbox(value = "extract json from this image.")
    submit = gr.Button("Submit")
    output = gr.TextArea()
    examples = gr.Examples(examples=EXAMPLES, inputs=[image])

    submit.click(blackjack_ai, inputs=[image, question], outputs=[output])

demo.launch()