app.py

import gradio as gr
import spaces
from transformers import Qwen2VLForConditionalGeneration, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch
from PIL import Image
import subprocess
from datetime import datetime
import numpy as np
import os


# subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)

# models = {
#     "Qwen/Qwen2-VL-2B-Instruct": AutoModelForCausalLM.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", trust_remote_code=True, torch_dtype="auto", _attn_implementation="flash_attention_2").cuda().eval()

# }
def array_to_image_path(image_array):
    # Convert numpy array to PIL Image
    img = Image.fromarray(np.uint8(image_array))
    
    # Generate a unique filename using timestamp
    # timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
    # filename = f"image_{timestamp}.png" # comment this, only save 1 image int the local path

    filename = "image_to_inference.png"
    
    # Save the image
    img.save(filename)
    
    # Get the full path of the saved image
    full_path = os.path.abspath(filename)
    
    return full_path
    
models = {
    "Qwen/Qwen2-VL-2B-Instruct": Qwen2VLForConditionalGeneration.from_pretrained(
        "Qwen/Qwen2-VL-2B-Instruct", 
        trust_remote_code=True, 
        # load_in_4bit=True,
        # attn_implementation="flash_attention_2",
        torch_dtype=torch.bfloat16
    ).cuda().eval()

}

processors = {
    "Qwen/Qwen2-VL-2B-Instruct": AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", trust_remote_code=True)
}

DESCRIPTION = "[Qwen2-VL-2B Demo](https://huggingface.co/Qwen/Qwen2-VL-2B-Instruct)"

kwargs = {}
kwargs['torch_dtype'] = torch.bfloat16

user_prompt = '<|user|>\n'
assistant_prompt = '<|assistant|>\n'
prompt_suffix = "<|end|>\n"

@spaces.GPU
def run_example(image, text_input=None, model_id="Qwen/Qwen2-VL-2B-Instruct"):
    image_path = array_to_image_path(image)
    
    print(image_path)
    model = models[model_id]
    processor = processors[model_id]

    prompt = f"{user_prompt}<|image_1|>\n{text_input}{prompt_suffix}{assistant_prompt}"
    image = Image.fromarray(image).convert("RGB")
    messages = [
    {
            "role": "user",
            "content": [
                {
                    "type": "image",
                    "image": image_path,
                },
                {"type": "text", "text": text_input},
            ],
        }
    ]
    
    # 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[0]

css = """
  #output {
    height: 500px; 
    overflow: auto; 
    border: 1px solid #ccc; 
  }
"""

with gr.Blocks(css=css) as demo:
    gr.Markdown(DESCRIPTION)
    with gr.Tab(label="Qwen2-VL-2B Input"):
        with gr.Row():
            with gr.Column():
                input_img = gr.Image(label="Input Picture")
                model_selector = gr.Dropdown(choices=list(models.keys()), label="Model", value="Qwen/Qwen2-VL-2B-Instruct")
                text_input = gr.Textbox(label="Question")
                submit_btn = gr.Button(value="Submit")
            with gr.Column():
                output_text = gr.Textbox(label="Output Text")

        submit_btn.click(run_example, [input_img, text_input, model_selector], [output_text])

demo.queue(api_open=False)
demo.launch(debug=True)