Error while adding mask_zero=True

[jerrychen1990]

when I add mask=zero in Embedding layer, CRF raises Error of types dismatch here is my code

and it raise exception like this: `--------------------------------------------------------------------------- ValueError Traceback (most recent call last) ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py in _apply_op_helper(self, op_type_name, name, **keywords) 454 preferred_dtype=default_dtype, --> 455 as_ref=input_arg.is_ref) 456 if input_arg.number_attr and len(

~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/tensorflow/python/framework/ops.py in internal_convert_n_to_tensor(values, dtype, name, as_ref, preferred_dtype, ctx) 1239 preferred_dtype=preferred_dtype, -> 1240 ctx=ctx)) 1241 return ret

~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/tensorflow/python/framework/ops.py in internal_convert_to_tensor(value, dtype, name, as_ref, preferred_dtype, ctx, accept_symbolic_tensors) 1174 if ret is None: -> 1175 ret = conversion_func(value, dtype=dtype, name=name, as_ref=as_ref) 1176

~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/tensorflow/python/framework/ops.py in _TensorTensorConversionFunction(t, dtype, name, as_ref) 976 "Tensor conversion requested dtype %s for Tensor with dtype %s: %r" % --> 977 (dtype.name, t.dtype.name, str(t))) 978 return t

ValueError: Tensor conversion requested dtype bool for Tensor with dtype float32: 'Tensor("crf_11/zeros_like_4:0", shape=(?, ?), dtype=float32)'

During handling of the above exception, another exception occurred:

TypeError Traceback (most recent call last)

in 5 6 crf = CRF(5,sparse_target=True) ----> 7 y = crf(emb) 8 model = Model(x, y) 9 model.summary() ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/keras/engine/base_layer.py in __call__(self, inputs, **kwargs) 448 if 'mask' not in kwargs: 449 # If mask is explicitly passed to __call__, --> 450 # we should override the default mask. 451 kwargs['mask'] = previous_mask 452 # Handle automatic shape inference (only useful for Theano). ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/keras_contrib/layers/crf.py in call(self, X, mask) 290 291 if self.test_mode == 'viterbi': --> 292 test_output = self.viterbi_decoding(X, mask) 293 else: 294 test_output = self.get_marginal_prob(X, mask) ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/keras_contrib/layers/crf.py in viterbi_decoding(self, X, mask) 562 input_energy, mask, self.left_boundary, self.right_boundary) 563 --> 564 argmin_tables = self.recursion(input_energy, mask, return_logZ=False) 565 argmin_tables = K.cast(argmin_tables, 'int32') 566 ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/keras_contrib/layers/crf.py in recursion(self, input_energy, mask, go_backwards, return_sequences, return_logZ, input_length) 514 515 if mask is not None: --> 516 mask2 = K.cast(K.concatenate([mask, K.zeros_like(mask[:, :1])], axis=1), 517 K.floatx()) 518 constants.append(mask2) ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py in concatenate(tensors, axis) 2162 if stop is None: 2163 try: -> 2164 if start < 0: 2165 start = 0 2166 except TypeError: ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/tensorflow/python/util/dispatch.py in wrapper(*args, **kwargs) 178 """Call target, and fall back on dispatchers if there is a TypeError.""" 179 try: --> 180 return target(*args, **kwargs) 181 except (TypeError, ValueError): 182 # Note: convert_to_eager_tensor currently raises a ValueError, not a ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/tensorflow/python/ops/array_ops.py in concat(values, axis, name) 1254 tensor_shape.scalar()) 1255 return identity(values[0], name=scope) -> 1256 return gen_array_ops.concat_v2(values=values, axis=axis, name=name) 1257 1258 ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/tensorflow/python/ops/gen_array_ops.py in concat_v2(values, axis, name) 1147 _attr_N = len(values) 1148 _, _, _op = _op_def_lib._apply_op_helper( -> 1149 "ConcatV2", values=values, axis=axis, name=name) 1150 _result = _op.outputs[:] 1151 _inputs_flat = _op.inputs ~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py in _apply_op_helper(self, op_type_name, name, **keywords) 481 (prefix, dtype.name)) 482 else: --> 483 raise TypeError("%s that don't all match." % prefix) 484 else: 485 raise TypeError( TypeError: Tensors in list passed to 'values' of 'ConcatV2' Op have types [bool, float32] that don't all match.`

[quancq]

I meet same problem when use keras_contrib 2.0.8, tf 1.12, keras 2.2.5. After install keras 2.2.4, this error disappear.

[DenceChen]

keras_contrib crf not support masking

[1535966643]

I meet same problem when use keras_contrib 2.0.8, tf 1.12, keras 2.2.5. After install keras 2.2.4, this error disappear.

hahahah,,, before have a keras bug when keras 2.2.4,so i install keras 2.2.5. now i meet this bug must need keras=2.2.4,,it's baddly

[FanYaning]

~/miniconda3/envs/chenhao-env/lib/python3.7/site-packages/keras_contrib/layers/crf.py in recursion(self, input_energy, mask, go_backwards, return_sequences, return_logZ, input_length)
514
515 if mask is not None:
--> 516 mask2 = K.cast(K.concatenate([mask, K.zeros_like(mask[:, :1])], axis=1),
517 K.floatx())
518 constants.append(mask2)

try to change mask2 to mask2 = K.concatenate([K.cast(mask, K.floatx()), K.zeros_like(mask[:, :1])], axis=1)