File size: 2,564 Bytes
0a82b18 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 |
import torch
from pathlib import Path
import gdown
import torchvision.transforms as tfm
from matching import WEIGHTS_DIR, THIRD_PARTY_DIR, BaseMatcher
from matching.utils import to_numpy, resize_to_divisible, lower_config, add_to_path
add_to_path(THIRD_PARTY_DIR.joinpath("MatchFormer"))
from model.matchformer import Matchformer
from config.defaultmf import get_cfg_defaults as mf_cfg_defaults
class MatchformerMatcher(BaseMatcher):
weights_src = "https://drive.google.com/file/d/1Ii-z3dwNwGaxoeFVSE44DqHdMhubYbQf/view"
weights_path = WEIGHTS_DIR.joinpath("matchformer_outdoor-large-LA.ckpt")
divisible_size = 32
def __init__(self, device="cpu", **kwargs):
super().__init__(device, **kwargs)
self.download_weights()
self.matcher = self.load_model().to(device).eval()
def download_weights(self):
if not Path(self.weights_path).is_file():
print("Downloading Matchformer outdoor... (takes a while)")
gdown.download(
MatchformerMatcher.weights_src,
output=str(self.weights_path),
fuzzy=True,
)
def load_model(self, cfg_path=None):
config = mf_cfg_defaults()
if cfg_path is not None:
config.merge_from_file(cfg_path)
config.MATCHFORMER.BACKBONE_TYPE = "largela"
config.MATCHFORMER.SCENS = "outdoor"
config.MATCHFORMER.RESOLUTION = (8, 2)
config.MATCHFORMER.COARSE.D_MODEL = 256
config.MATCHFORMER.COARSE.D_FFN = 256
matcher = Matchformer(config=lower_config(config)["matchformer"])
matcher.load_state_dict(
{k.replace("matcher.", ""): v for k, v in torch.load(self.weights_path, map_location="cpu").items()}
)
return matcher
def preprocess(self, img):
_, h, w = img.shape
orig_shape = h, w
img = resize_to_divisible(img, self.divisible_size)
return tfm.Grayscale()(img).unsqueeze(0), orig_shape
def _forward(self, img0, img1):
img0, img0_orig_shape = self.preprocess(img0)
img1, img1_orig_shape = self.preprocess(img1)
batch = {"image0": img0, "image1": img1}
self.matcher(batch)
mkpts0 = to_numpy(batch["mkpts0_f"])
mkpts1 = to_numpy(batch["mkpts1_f"])
H0, W0, H1, W1 = *img0.shape[-2:], *img1.shape[-2:]
mkpts0 = self.rescale_coords(mkpts0, *img0_orig_shape, H0, W0)
mkpts1 = self.rescale_coords(mkpts1, *img1_orig_shape, H1, W1)
return mkpts0, mkpts1, None, None, None, None
|