app.py

import torch
from transformers import AutoProcessor, AutoModelForVision2Seq
from fastapi import FastAPI, File, UploadFile, Form
from fastapi.responses import JSONResponse
from io import BytesIO
from PIL import Image

model_id = "HPAI-BSC/Aloe-Vision-7B-AR"

processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForVision2Seq.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True,
)

app = FastAPI(title="Aloe Vision 7B AR API")

@app.post("/predict")
async def predict(
    file: UploadFile = File(None),
    question: str = Form(None)
):
    try:
        # --- Case 1: both image and text ---
        if file and question:
            image = Image.open(BytesIO(await file.read())).convert("RGB")
            messages = [
                {"role": "user", "content": [
                    {"type": "image", "image": image},
                    {"type": "text", "text": question}
                ]}
            ]

        # --- Case 2: text only ---
        elif question and not file:
            messages = [{"role": "user", "content": [{"type": "text", "text": question}]}]

        # --- Case 3: image only ---
        elif file and not question:
            image = Image.open(BytesIO(await file.read())).convert("RGB")
            messages = [
                {"role": "user", "content": [
                    {"type": "image", "image": image},
                    {"type": "text", "text": "Describe this image briefly."}
                ]}
            ]
        else:
            return JSONResponse({"error": "You must provide an image, text, or both."}, status_code=400)

        # --- Process ---
        text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
        image_inputs = processor.process_vision_info(messages)
        inputs = processor(text=[text], **image_inputs, return_tensors="pt").to(model.device)

        generated = model.generate(
            **inputs,
            max_new_tokens=256,
            do_sample=False,
            eos_token_id=processor.tokenizer.eos_token_id,
        )

        output_text = processor.batch_decode(generated, skip_special_tokens=True)[0]
        answer = output_text.split(text)[-1].strip()

        return JSONResponse({"answer": answer})

    except Exception as e:
        return JSONResponse({"error": str(e)}, status_code=500)