[GitHub] soeque1 commented on issue #12844: ERROR parameter summary of lstm layer

[GitBox <gi...@apache.org>]

soeque1 commented on issue #12844: ERROR parameter summary of lstm layer
URL: https://github.com/apache/incubator-mxnet/issues/12844#issuecomment-430985305
 
 
   Thanks for your help in solving this problem. 
   
   But, There are still some issues. This is not HybridSequential or Sequential.
   
   ```
   import mxnet as mx
   from mxnet import gluon
   from mxnet.gluon import HybridBlock, nn, rnn
   class MyModel(gluon.HybridBlock):
       def __init__(self, vocab_size, num_embed, **kwargs):
           super(MyModel, self).__init__(**kwargs)
           with self.name_scope():
               self.embed = nn.Embedding(input_dim=vocab_size, output_dim=num_embed)
               self.lstm = rnn.LSTM(20)
               self.out = nn.Dense(2)  
       def hybrid_forward(self, F ,inputs):
           em_out = self.embed(inputs)
           lstm_out = self.lstm(em_out) 
           return(self.out(lstm_out))
   
   model = MyModel(vocab_size =20, num_embed=50)
   
   model.initialize(mx.init.Xavier())
   model.hybridize()
   
   mx.viz.print_summary(
       model(mx.sym.var('data')), 
       shape={'data':(1,30)}, #set your shape here
   )
   
   ```
   ________________________________________________________________________________________________________________________
   Layer (type)                                        Output Shape            Param #     Previous Layer                  
   ========================================================================================================================
   data(null)                                          30                      0                                           
   ________________________________________________________________________________________________________________________
   mymodel3_embedding0_fwd(Embedding)                  30x50                   0           data                            
   ________________________________________________________________________________________________________________________
   mymodel3_lstm0_reshape0(Reshape)                                            0                                           
   ________________________________________________________________________________________________________________________
   mymodel3_lstm0_reshape1(Reshape)                                            0                                           
   ________________________________________________________________________________________________________________________
   mymodel3_lstm0_reshape2(Reshape)                                            0                                           
   ________________________________________________________________________________________________________________________
   mymodel3_lstm0_reshape3(Reshape)                                            0                                           
   ________________________________________________________________________________________________________________________
   
   ---------------------------------------------------------------------------
   IndexError                                Traceback (most recent call last)
   <ipython-input-8-d83349ac28a7> in <module>
        21 mx.viz.print_summary(
        22     model(mx.sym.var('data')),
   ---> 23     shape={'data':(1,30)}, #set your shape here
        24 )
   
   ~/python/venv/ngram_detectors/lib/python3.6/site-packages/mxnet/visualization.py in print_summary(symbol, shape, line_length, positions)
       186                 if key in shape_dict:
       187                     out_shape = shape_dict[key][1:]
   --> 188         total_params += print_layer_summary(nodes[i], out_shape)
       189         if i == len(nodes) - 1:
       190             print('=' * line_length)
   
   ~/python/venv/ngram_detectors/lib/python3.6/site-packages/mxnet/visualization.py in print_layer_summary(node, out_shape)
       132                         if key in shape_dict:
       133                             shape = shape_dict[key][1:]
   --> 134                             pre_filter = pre_filter + int(shape[0])
       135         cur_param = 0
       136         if op == 'Convolution':
   
   IndexError: tuple index out of range
   
   ```
   
   
   
   
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services