app.py

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

# Default: Load the model on the available device(s)
model = Qwen2VLForConditionalGeneration.from_pretrained(
    "OS-Copilot/OS-Atlas-Base-7B", torch_dtype="auto", device_map="auto"
)
processor = AutoProcessor.from_pretrained("OS-Copilot/OS-Atlas-Base-7B")
@spaces.GPU
def run(image, message):
    messages = [
        {
            "role": "user",
            "content": [
                {
                    "type": "image",
                    "image": "image,
                },
                {"type": "text", "text": message},
            ],
        }
    ]
    
    
    # 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=False, clean_up_tokenization_spaces=False
    )
    return output_text
    # <|object_ref_start|>language switch<|object_ref_end|><|box_start|>(576,12),(592,42)<|box_end|><|im_end|>

with gr.Blocks() as demo:
    gr.Markdown("# Unofficial OS-Atlas demo")
    image = gr.Image(label="Image")
    text = gr.Textbox(label="Prompt")
    btn = gr.Button("Generate", variant="primary")
    output = gr.Textbox(interactive=False)
    btn.click(run, inputs=[image, text], outputs=output)
demo.queue().launch()