• 版本、环境信息：

   1）PaddlePaddle版本：2.0rc
   2）环境：AI Studio高级版

paddle.vision.models.VGG() 的参数 num_classes 介绍写道：最后一个全连接层输出的维度。如果该值小于0，则不定义最后一个全连接层。默认值：1000。但是测试的时候发现如果该值小于等于0，都会报错

##### 示例代码 #####

import paddle
from paddle.vision.models import VGG
from paddle.vision.models.vgg import make_layers

vgg11_cfg = [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M']

features = make_layers(vgg11_cfg)

vgg11 = VGG(features, num_classes=0)

x = paddle.rand([1, 3, 224, 224])
out = vgg11(x)

print(out.shape)

##### 错误信息  #####

---------------------------------------------------------------------------EnforceNotMet                             Traceback (most recent call last)<ipython-input-69-4ae249ce9371> in <module>
     10 
     11 x = paddle.rand([1, 3, 224, 224])
---> 12 out = vgg11(x)
     13 
     14 print(out.shape)
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py in __call__(self, *inputs,**kwargs)
    827 
    828         with param_guard(self._parameters), param_guard(self._buffers):
--> 829             outputs = self.forward(*inputs,**kwargs)
    830 
    831         for forward_post_hook in self._forward_post_hooks.values():
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/vision/models/vgg.py in forward(self, x)
     72         x = self.avgpool(x)
     73         x = paddle.flatten(x, 1)
---> 74         x = self.classifier(x)
     75         return x
     76 
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py in __call__(self, *inputs,**kwargs)
    827 
    828         with param_guard(self._parameters), param_guard(self._buffers):
--> 829             outputs = self.forward(*inputs,**kwargs)
    830 
    831         for forward_post_hook in self._forward_post_hooks.values():
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/container.py in forward(self, input)
     84     def forward(self, input):
     85         for layer in self._sub_layers.values():
---> 86             input = layer(input)
     87         return input
     88 
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py in __call__(self, *inputs,**kwargs)
    827 
    828         with param_guard(self._parameters), param_guard(self._buffers):
--> 829             outputs = self.forward(*inputs,**kwargs)
    830 
    831         for forward_post_hook in self._forward_post_hooks.values():
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/nn/layer/common.py in forward(self, input)
    137     def forward(self, input):
    138         out = F.linear(
--> 139             x=input, weight=self.weight, bias=self.bias, name=self.name)
    140         return out
    141 
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/nn/functional/common.py in linear(x, weight, bias, name)
   1454         pre_bias = _varbase_creator(dtype=x.dtype)
   1455         core.ops.matmul(x, weight, pre_bias, 'transpose_X', False,
-> 1456                         'transpose_Y', False, "alpha", 1)
   1457         return dygraph_utils._append_bias_in_dygraph(
   1458             pre_bias, bias, axis=len(x.shape) - 1)
EnforceNotMet: 
----------------------
Error Message Summary:
----------------------
ExternalError:  Cublas error, CUBLAS_STATUS_INVALID_VALUE  (at /paddle/paddle/fluid/operators/math/blas_impl.cu.h:34)
  [operator < matmul > error]

还有一点想说的是，文档的描述和API内置文档的描述不一样。比如文档写的是 ”如果该值小于0，则不定义最后一个全连接层。“ ，但help()函数里面的描述是 ”If num_classes <=0, last fc layer will not be defined.“