Slow LightGlue at 1,7s

[ghost]

"So, I'm working with large images (5472x3648), and my code is taking around 1.7 seconds. Can someone help? LightGlue is supposed to operate in milliseconds."

from pathlib import Path
from lightglue import LightGlue, SuperPoint
from lightglue.utils import load_image, rbd
import torch
import cv2
import numpy as np
import time

torch.set_grad_enabled(False)
images = Path("C:\Users\irh\Documents\img_alinhamento\1\")

device = torch.device("cuda") # 'mps', 'cpu'

extractor = SuperPoint(max_num_keypoints=2048).eval().to(device) # load the extractor
matcher = LightGlue(features="superpoint").eval().to(device)
a= time.time()
image0 = load_image(images / "cabo_fan_dir_ok(2).jpg")
image1 = load_image(images / "bateria_nok(2).jpg")
img1 = cv2.imread('C:\Users\irh\Documents\img_alinhamento\1\cabo_fan_dir_ok(2).jpg')
with torch.no_grad():
feats0 = extractor.extract(image0.to(device))
feats1 = extractor.extract(image1.to(device))
matches01 = matcher({"image0": feats0, "image1": feats1})
feats0, feats1, matches01 = [
rbd(x) for x in [feats0, feats1, matches01]
] # remove batch dimension

kpts0, kpts1, matches = feats0["keypoints"], feats1["keypoints"], matches01["matches"]
m_kpts0, m_kpts1 = kpts0[matches[..., 0]], kpts1[matches[..., 1]]
m_kpts0, m_kpts1 = m_kpts0.cpu().numpy(), m_kpts1.cpu().numpy()
b= time.time()
c = b- a
print(c)
M, status = cv2.estimateAffinePartial2D(np.float32(m_kpts0),
np.float32(m_kpts1), cv2.RANSAC)
result = cv2.warpAffine(img1, M, (img1.shape[1], img1.shape[0]),cv2.INTER_NEAREST_EXACT)
cv2.imwrite('C:\Users\irh\Documents\img_alinhamento\1\alinhada.jpg',result)