def predict(model):
# 读入模型
model = load_checkpoint(model)
print('..... Finished loading model! ......')
##将模型放置在gpu上运行
if torch.cuda.is_available():
model.cuda()
pred_list, _id, pred_pro_list = [], [], []
for i in tqdm(range(len(imgs))):
img_path = imgs[i].strip()
# print(img_path)
_id.append(os.path.basename(img_path).split('.')[0])
img = Image.open(img_path).convert('RGB')
# print(type(img))
img = get_test_transform(size=cfg.INPUT_SIZE)(img).unsqueeze(0)
if torch.cuda.is_available():
img = img.cuda()
with torch.no_grad():
out = model(img)
softmax = F.softmax(out, dim=-1)
#print(softmax.tolist()[0])
pred_pro_list.append(softmax.tolist()[0])
prediction = torch.argmax(out, dim=1).cpu().item()
pred_list.append(prediction)
return _id, pred_list, pred_pro_list
def __init__(self, label_file, imageset):
'''
img_dir: 图片路径：img_dir + img_name.jpg构成图片的完整路径
'''
# 所有图片的绝对路径
with open(label_file, 'r') as f:
#label_file的格式， （label_file image_label)
self.imgs = list(map(lambda line: line.strip().split(' '), f))
# 相关预处理的初始化
# self.transforms=transform
self.img_aug=True
if imageset == 'train':
self.transform= get_train_transform(size=cfg.INPUT_SIZE)
else:
self.transform = get_test_transform(size = cfg.INPUT_SIZE)
self.input_size = cfg.INPUT_SIZE
def predict(model):
# 读入模型
model = load_checkpoint(model)
print('..... Finished loading model! ......')
##将模型放置在gpu上运行
if torch.cuda.is_available():
model.cuda()
img = Image.open(img_path).convert('RGB')
img = get_test_transform(size=cfg.INPUT_SIZE)(img).unsqueeze(0)
if torch.cuda.is_available():
img = img.cuda()
with torch.no_grad():
x = model.conv1(img)
draw_features(8, 8, 64, x.cpu().numpy(), "{}/f1_conv1.png".format(savepath))
def __init__(self, label_file, imageset):
'''
img_dir: 图片路径：img_dir + img_name.jpg构成图片的完整路径
'''
# 所有图片的绝对路径
label_file = "D:\\05分类图片\\pytorch_classification-master\\data\\train.txt"
print(label_file)
with open(label_file, 'r') as f:
#label_file的格式， （label_file image_label)
self.imgs = list(map(lambda line: line.strip().split(' '), f))
# 相关预处理的初始化
# self.transforms=transform
self.img_aug=True
if imageset == 'train':
self.transform= get_train_transform(size=cfg.INPUT_SIZE)
else:
self.transform = get_test_transform(size = cfg.INPUT_SIZE)
self.eraser = get_random_eraser( s_h=0.1, pixel_level=True)
self.input_size = cfg.INPUT_SIZE