#load model

from unsloth import FastVisionModel
import torch

model, tokenizer = FastVisionModel.from_pretrained(
    "manifestasi/Manifestasi-Qwen-VL-2b-ft",
    load_in_4bit=True, 
    use_gradient_checkpointing="unsloth"
)
FastVisionModel.for_inference(model)

#run inference

from time import time

t1 = time()

#image kamu
image = img_resized
instruction = "apa tulisan ini ?"


messages = [
    {"role": "user", "content": [
        {"type": "image"},
        {"type": "text", "text": instruction}
    ]}
]
input_text = tokenizer.apply_chat_template(messages, add_generation_prompt=True)


inputs = tokenizer(
    image,
    input_text,
    add_special_tokens=False,
    return_tensors="pt"
).to("cuda")


generated_ids = model.generate(
    **inputs,
    max_new_tokens=128,
    use_cache=True,
    temperature=1.5,
    min_p=0.1
)

#Decode token yang dihasilkan menjadi string
output_text = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
t2 = time()

#Simpan output ke dalam string dan cetak hasilnya
print("Output text:", output_text)
print("waktu : ", (t2-t1))