running error "movielens-learning.ipynb" in [11] Init mode #2
Description
Hello,
I tried to run "movielens-learning.ipynb" and I got this following error (in [11], "Init mode" part):
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-11-b8fe01c64add> in <module>()
3 model = TFExpMachine(rank=target_rank, s_features=s_features, init_std=0.001, reg=0.012, exp_reg=1.8)
4 model.init_from_cores(w_cores)
----> 5 model.build_graph()
6 model.initialize_session()
/home/michele/exp-machines/src/TFExpMachine.pyc in build_graph(self)
62 cur_col = self.X[:, 0]
63 tower = tf.gather(self.G[0], cur_col)
---> 64 self.outputs = tf.add(self.G[0][0], tower)
65 for i in range(1, self.n_features):
66 cur_col = self.X[:, i]
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/ops/variables.pyc in <lambda>(a, b)
613 setattr(Variable, operator, lambda a: Variable._RunOp(operator, a, None))
614 else:
--> 615 setattr(Variable, operator, lambda a, b: Variable._RunOp(operator, a, b))
616
617 # NOTE(mrry): This enables the Variable's overloaded "right" binary
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/ops/variables.pyc in _RunOp(operator, a, b)
637 # pylint: disable=protected-access
638 if b is not None:
--> 639 return getattr(ops.Tensor, operator)(a._AsTensor(), b)
640 else:
641 return getattr(ops.Tensor, operator)(a._AsTensor())
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/ops/array_ops.pyc in _SliceHelper(tensor, slice_spec)
336 sizes.append(1)
337 squeeze_dims.append(dim)
--> 338 sliced = slice(tensor, indices, sizes)
339 if squeeze_dims:
340 return squeeze(sliced, squeeze_dims=squeeze_dims)
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/ops/array_ops.pyc in slice(input_, begin, size, name)
386 A `Tensor` the same type as `input`.
387 """
--> 388 return gen_array_ops._slice(input_, begin, size, name=name)
389
390
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/ops/gen_array_ops.pyc in _slice(input, begin, size, name)
1999 """
2000 result = _op_def_lib.apply_op("Slice", input=input, begin=begin, size=size,
-> 2001 name=name)
2002 return result
2003
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/framework/op_def_library.pyc in apply_op(self, op_type_name, name, **keywords)
701 op = g.create_op(op_type_name, inputs, output_types, name=scope,
702 input_types=input_types, attrs=attr_protos,
--> 703 op_def=op_def)
704 outputs = op.outputs
705 return _Restructure(ops.convert_n_to_tensor(outputs),
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/framework/ops.pyc in create_op(self, op_type, inputs, dtypes, input_types, name, attrs, op_def, compute_shapes, compute_device)
2310 original_op=self._default_original_op, op_def=op_def)
2311 if compute_shapes:
-> 2312 set_shapes_for_outputs(ret)
2313 self._add_op(ret)
2314 self._record_op_seen_by_control_dependencies(ret)
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/framework/ops.pyc in set_shapes_for_outputs(op)
1702 raise RuntimeError("No shape function registered for standard op: %s"
1703 % op.type)
-> 1704 shapes = shape_func(op)
1705 if shapes is None:
1706 raise RuntimeError(
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/ops/array_ops.pyc in _SliceShape(op)
1508 ndims = begin_shape.merge_with(sizes_shape)[0].value
1509 if ndims is not None:
-> 1510 input_shape.assert_has_rank(ndims)
1511 begin_value = tensor_util.constant_value(op.inputs[1])
1512 sizes_value = tensor_util.constant_value(op.inputs[2])
/home/michele/miniconda2/lib/python2.7/site-packages/tensorflow/python/framework/tensor_shape.pyc in assert_has_rank(self, rank)
619 """
620 if self.ndims not in (None, rank):
--> 621 raise ValueError("Shape %s must have rank %d" % (self, rank))
622
623 def with_rank(self, rank):
ValueError: Shape (8, 1, 10) must have rank 1
Can you please help me with this above error ?