Created by: xfliu138

操作系统win7 64位，python版本是3.7.64，本机测试图像识别示例程序image_classification。 训练完成后，用image目录下的dog.png图片验证是正确的，但从网络上随机找了一些猫，船之类的图片，验证时提示如下错误：

Traceback (most recent call last): File "train.py", line 209, in main(use_cuda=False) File "train.py", line 203, in main infer(use_cuda=use_cuda, params_dirname=save_path) File "train.py", line 175, in infer fetch_list=fetch_targets) File "C:\Users\miles.liu\AppData\Local\Programs\Python\Python37\lib\site-packa ges\paddle\fluid\executor.py", line 472, in run self.executor.run(program.desc, scope, 0, True, True) paddle.fluid.core.EnforceNotMet: Enforce failed. Expected in_dims[1] == filter_d ims[1] * groups, but received in_dims[1]:4 != filter_dims[1] * groups:3. The number of input channels should be equal to filter channels * groups. at [E: \dist\Paddle\paddle\fluid\operators\conv_op.cc:60] PaddlePaddle Call Stacks: Windows not support stack backtrace yet. 开始以为图片格式有问题，然后将图片格式改为png，长和高都设定为32与dog.png一致，但还是报以上错误