Created by: rationme

'EnforceNotMet: enforce in_dims.size() == filter_dims.size() failed, 4 != 1 Conv input dimension and filter dimension should be the same. at [/paddle/paddle/fluid/operators/conv_op.cc:49]'

when i run my code main_program, feed=feeder.feed(data),fetch_list=[l2_cost]) and `def CSRNet(input,label): def conv_block(ipt, filter_list,pool_size = 2): return fluid.nets.img_conv_group( input=ipt, pool_size=pool_size, pool_stride=pool_size, conv_num_filter=filter_list, conv_filter_size=3, conv_act='relu', conv_with_batchnorm=False, conv_batchnorm_drop_rate=1, pool_type='max')

frontend_conv1 = conv_block(input,[64,64])

frontend_conv2 = conv_block(frontend_conv1,[128,128])

frontend_conv3 = conv_block(frontend_conv2 ,[256, 256,256])

frontend_conv4 = conv_block(frontend_conv3,[512,512,512],1)`

I can't find where the input dimension is?how to use it?