File size: 3,128 Bytes
bdec3d7 |
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 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 |
# --------------------------------------------------------
# X-Decoder -- Generalized Decoding for Pixel, Image, and Language
# Copyright (c) 2022 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# Written by Xueyan Zou ([email protected])
# --------------------------------------------------------
import os
import sys
import logging
pth = '/'.join(sys.path[0].split('/')[:-1])
sys.path.insert(0, pth)
from PIL import Image
import numpy as np
np.random.seed(0)
import cv2
import torch
from torchvision import transforms
from utils.arguments import load_opt_command
from detectron2.data import MetadataCatalog
from detectron2.structures import BitMasks
from modeling.BaseModel import BaseModel
from modeling import build_model
from detectron2.utils.colormap import random_color
from utils.visualizer import Visualizer
from utils.distributed import init_distributed
logger = logging.getLogger(__name__)
def main(args=None):
'''
Main execution point for PyLearn.
'''
opt, cmdline_args = load_opt_command(args)
if cmdline_args.user_dir:
absolute_user_dir = os.path.abspath(cmdline_args.user_dir)
opt['base_path'] = absolute_user_dir
opt = init_distributed(opt)
# META DATA
pretrained_pth = os.path.join(opt['RESUME_FROM'])
if 'novg' not in pretrained_pth:
assert False, "Using the ckpt without visual genome training data will be much better."
output_root = './output'
image_pth = 'inference/images/mountain.jpeg'
model = BaseModel(opt, build_model(opt)).from_pretrained(pretrained_pth).eval().cuda()
model.model.sem_seg_head.predictor.lang_encoder.get_text_embeddings(["background"], is_eval=False)
t = []
t.append(transforms.Resize(224, interpolation=Image.BICUBIC))
transform = transforms.Compose(t)
with torch.no_grad():
image_ori = Image.open(image_pth).convert("RGB")
width = image_ori.size[0]
height = image_ori.size[1]
image = transform(image_ori)
image = np.asarray(image)
image_ori = np.asarray(image_ori)
images = torch.from_numpy(image.copy()).permute(2,0,1).cuda()
batch_inputs = [{'image': images, 'height': height, 'width': width, 'image_id': 0}]
outputs = model.model.evaluate_captioning(batch_inputs)
text = outputs[-1]['captioning_text']
image_ori = image_ori[:,:,::-1].copy()
cv2.rectangle(image_ori, (0, 0), (width, 60), (0,0,0), -1)
font = cv2.FONT_HERSHEY_DUPLEX
fontScale = 1.2
thickness = 2
lineType = 2
bottomLeftCornerOfText = (10, 40)
fontColor = [255,255,255]
cv2.putText(image_ori, text,
bottomLeftCornerOfText,
font,
fontScale,
fontColor,
thickness,
lineType)
if not os.path.exists(output_root):
os.makedirs(output_root)
cv2.imwrite(os.path.join(output_root, 'captioning.png'), image_ori)
if __name__ == "__main__":
main()
sys.exit(0) |