def crnnSource(net_path, alphabet):
converter = util.strLabelConverter(alphabet)
if cfg.PLATFORM == "GPU":
model = crnn.CRNN(32, 1, len(alphabet) + 1, 256, 1).cuda()
#model = torch.nn.DataParallel(model, device_ids=range(1))
model.load_state_dict(torch.load(net_path))
else:
model = crnn.CRNN(32, 1, len(alphabet) + 1, 256, 1)
model.load_state_dict(torch.load(net_path, map_location=lambda storage, loc: storage))
return model,converter
def crnnSource(net_path, alphabet):
converter = util.strLabelConverter(alphabet)
if config.PLATFORM == "GPU":
model = crnn.CRNN(32, 1, len(alphabet) + 1, 256, 1).cuda()
model.load_state_dict(torch.load(net_path))
else:
model = crnn.CRNN(32, 1, len(alphabet) + 1, 256, 1)
model.load_state_dict(
torch.load(net_path, map_location=lambda storage, loc: storage))
#model.load_state_dict(torch.load(net_path))
return model, converter
def __init__(self, model_path=CONFIG.model_path, alphabet=CONFIG.alphabet):
self.model = crnn.CRNN(n_class=len(alphabet) + 1)
self.model.load_state_dict(torch.load(model_path)['state_dict'])
self.converter = util.strLabelConverter(alphabet)
self.toTensor = transforms.ToTensor()
self.model = self.model.cuda()
self.model.eval()
# test if crnn work
if __name__ == '__main__':
imagepath = './test.jpg'
img_h = opt.img_h
use_gpu = opt.use_gpu
modelpath = opt.modelpath
char_set = open('char_std_5990.txt', 'r', encoding='utf-8').readlines()
char_set = ''.join([ch.strip('\n') for ch in char_set[1:]] + ['卍'])
n_class = len(char_set)
model = crnn.CRNN(img_h, 1, n_class, 256)
if torch.cuda.is_available and use_gpu:
model.cuda()
if os.path.exists(modelpath):
print('Load model from "%s" ...' % modelpath)
model.load_state_dict(torch.load(modelpath))
print('Done!')
image = Image.open(imagepath).convert('L')
(w, h) = image.size
size_h = 32
ratio = size_h / float(h)
size_w = int(w * ratio)
# keep the ratio
transform = resizeNormalize((size_w, size_h))