在完成 GoogLeNet 模型训练后,读取参数出现问题

读取其他模型的参数没有问题，但GoogLeNet 出现问题：Traceback (most recent call last): File "D:/python_adver/attack_code/acc.py", line 135, in <module> out = fluid.layers.softmax(out_logits) File "D:\python_adver\attack_code\venv\lib\site-packages\paddle\fluid\layers\nn.py", line 1983, in softmax "use_cudnn": use_cudnn}) File "D:\python_adver\attack_code\venv\lib\site-packages\paddle\fluid\layer_helper.py", line 43, in append_op return self.main_program.current_block().append_op(*args, **kwargs) File "D:\python_adver\attack_code\venv\lib\site-packages\paddle\fluid\framework.py", line 1748, in append_op attrs=kwargs.get("attrs", None) File "D:\python_adver\attack_code\venv\lib\site-packages\paddle\fluid\framework.py", line 1095, in __init__ "not suprt args type , should be[ string_type, binary_type, Varibale]" ValueError: not suprt args type , should be[ string_type, binary_type, Varibale]