SNN_RBM shape mismatch

[CorpaciLC]

Hi,

When running SNN_RBM.py on the datasets from the paper (tried on 1458 and 2261), we get the following error:

''' ../../make-ipinyou-data/2261/train.yzx.txt drop_mlp4da.py|ad:2261|drop:1|b_size:1000 | X:133465 | Hidden 0:300 | Hidden 1:300 | Hidden 2:100 | L_r:0.0006 | activation1:tanh | lambda:0.0001 Traceback (most recent call last): File "C:\Users\corpa\anaconda3\lib\site-packages\theano\compile\function_module.py", line 903, in call self.fn() if output_subset is None else
ValueError: Shape mismatch: x has 37 cols (and 1 rows) but y has 70 rows (and 300 cols)

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "SNN_DAE.py", line 85, in ww0,bb0,ww1,bb1,ww2,bb2=da.get_da_weights(train_file,arr,num_feats=numf,ncases=train_size,batch_size=100000) File "C:\Users\corpa\Master\Sem3\Experiment Design For Data Science\Assignment2\deep-ctr\python\sampling_based_denosing_autoencoder.py", line 364, in get_da_weights w,b=sparse_da(num_feats*k,col,file,training_epochs=epochs,sparse_len=row,is_sparse=1,batch_size=1,k=k) File "C:\Users\corpa\Master\Sem3\Experiment Design For Data Science\Assignment2\deep-ctr\python\sampling_based_denosing_autoencoder.py", line 327, in sparse_da [cost,z,w,b]=train_da(batcharr,initial_W) File "C:\Users\corpa\anaconda3\lib\site-packages\theano\compile\function_module.py", line 914, in call gof.link.raise_with_op( File "C:\Users\corpa\anaconda3\lib\site-packages\theano\gof\link.py", line 325, in raise_with_op reraise(exc_type, exc_value, exc_trace) File "C:\Users\corpa\anaconda3\lib\site-packages\six.py", line 702, in reraise raise value.with_traceback(tb) File "C:\Users\corpa\anaconda3\lib\site-packages\theano\compile\function_module.py", line 903, in call self.fn() if output_subset is None else
ValueError: Shape mismatch: x has 37 cols (and 1 rows) but y has 70 rows (and 300 cols) Apply node that caused the error: Dot22(Elemwise{Mul}[(0, 0)].0, ww) Toposort index: 12 Inputs types: [TensorType(float64, matrix), TensorType(float64, matrix)] Inputs shapes: [(1, 37), (70, 300)] Inputs strides: [(296, 8), (2400, 8)] Inputs values: ['not shown', 'not shown'] Outputs clients: [[Elemwise{Composite{scalar_sigmoid((i0 + i1))}}[(0, 0)](Dot22.0, InplaceDimShuffle{x,0}.0)]]

Backtrace when the node is created(use Theano flag traceback.limit=N to make it longer): File "SNN_DAE.py", line 85, in ww0,bb0,ww1,bb1,ww2,bb2=da.get_da_weights(train_file,arr,num_feats=numf,ncases=train_size,batch_size=100000) File "C:\Users\corpa\Master\Sem3\Experiment Design For Data Science\Assignment2\deep-ctr\python\sampling_based_denosing_autoencoder.py", line 364, in get_da_weights w,b=sparse_da(num_feats*k,col,file,training_epochs=epochs,sparse_len=row,is_sparse=1,batch_size=1,k=k) File "C:\Users\corpa\Master\Sem3\Experiment Design For Data Science\Assignment2\deep-ctr\python\sampling_based_denosing_autoencoder.py", line 249, in sparse_da cost, updates,z,w,b = da.get_cost_updates( File "C:\Users\corpa\Master\Sem3\Experiment Design For Data Science\Assignment2\deep-ctr\python\sampling_based_denosing_autoencoder.py", line 102, in get_cost_updates y = self.get_hidden_values(tilde_x) File "C:\Users\corpa\Master\Sem3\Experiment Design For Data Science\Assignment2\deep-ctr\python\sampling_based_denosing_autoencoder.py", line 92, in get_hidden_values return T.nnet.sigmoid(T.dot(input, self.W) + self.b)

HINT: Use the Theano flag 'exception_verbosity=high' for a debugprint and storage map footprint of this apply node. '''

Ideas of solutions?

Thanks!