app.py

images_dir = "images"
import io
from transformers import Qwen2VLForConditionalGeneration, AutoProcessor
from qwen_vl_utils import process_vision_info
from PIL import Image
import torch
torch.cuda.empty_cache()

from fastapi import FastAPI, File, Form,UploadFile,HTTPException

app=FastAPI()

def run_model(image,text_input):
    torch.cuda.empty_cache()
    model_id= "Qwen/Qwen2-VL-7B-Instruct-AWQ"
    model = Qwen2VLForConditionalGeneration.from_pretrained(
        model_id , torch_dtype=torch.float16, device_map="cuda:0"
    )
    min_pixels = 256*28*28
    max_pixels = 1280*28*28
    processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-7B-Instruct-AWQ", min_pixels=min_pixels, max_pixels=max_pixels)

    torch.cuda.empty_cache()
    image_path = Image.open(image)
    print(image_path)
    messages = [
    {
            "role": "user",
            "content": [
                {
                    "type": "image",
                    "image": image_path,
                },
                {"type": "text", "text": text_input},
            ],
        }
    ]
   
    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
    torch.cuda.empty_cache()

    generated_ids = model.generate(**inputs, max_new_tokens=1024)
    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]


@app.post("/call_qwen_model")
async def call_model(file: UploadFile = File(...),json_str: str = Form(...)):
    try:
        request_object_content = await file.read()
        img = io.BytesIO(request_object_content)
        output = run_model(img, json_str)
        return  {"output": output}
    except Exception as e :
        raise HTTPException (f"Error: {e}")