InvalidArgumentError: Incompatible shapes running examples with keras 2.2.3 and later

[SunYanCN]

I do nothing,just run sequential_example.py. Then get a error as:

Using TensorFlow backend.
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
embedding_1 (Embedding)      (None, None, 10)          500       
_________________________________________________________________
bidirectional_1 (Bidirection (None, None, 300)         193200    
_________________________________________________________________
AttentionDecoder (AttentionD (None, None, 50)          343810    
=================================================================
Total params: 537,510
Trainable params: 537,510
Non-trainable params: 0
_________________________________________________________________
/root/anaconda3/lib/python3.6/site-packages/tensorflow/python/ops/gradients_impl.py:100: UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory.
  "Converting sparse IndexedSlices to a dense Tensor of unknown shape. "
Epoch 1/10
2019-01-23 14:11:08.400572: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2019-01-23 14:11:08.510241: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:897] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-01-23 14:11:08.510750: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1392] Found device 0 with properties: 
name: Tesla P4 major: 6 minor: 1 memoryClockRate(GHz): 1.1135
pciBusID: 0000:00:06.0
totalMemory: 7.43GiB freeMemory: 7.31GiB
2019-01-23 14:11:08.510791: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1471] Adding visible gpu devices: 0
2019-01-23 14:11:08.851793: I tensorflow/core/common_runtime/gpu/gpu_device.cc:952] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-01-23 14:11:08.851847: I tensorflow/core/common_runtime/gpu/gpu_device.cc:958]      0 
2019-01-23 14:11:08.851856: I tensorflow/core/common_runtime/gpu/gpu_device.cc:971] 0:   N 
2019-01-23 14:11:08.852059: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1084] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 7057 MB memory) -> physical GPU (device: 0, name: Tesla P4, pci bus id: 0000:00:06.0, compute capability: 6.1)
Traceback (most recent call last):
  File "/home/sunyan/JDDC/keras-monotonic-attention/sequential_example.py", line 33, in <module>
    model.fit(x, y, epochs=10)
  File "/root/anaconda3/lib/python3.6/site-packages/keras/engine/training.py", line 1039, in fit
    validation_steps=validation_steps)
  File "/root/anaconda3/lib/python3.6/site-packages/keras/engine/training_arrays.py", line 199, in fit_loop
    outs = f(ins_batch)
  File "/root/anaconda3/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py", line 2715, in __call__
    return self._call(inputs)
  File "/root/anaconda3/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py", line 2675, in _call
    fetched = self._callable_fn(*array_vals)
  File "/root/anaconda3/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1454, in __call__
    self._session._session, self._handle, args, status, None)
  File "/root/anaconda3/lib/python3.6/site-packages/tensorflow/python/framework/errors_impl.py", line 519, in __exit__
    c_api.TF_GetCode(self.status.status))
tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [320] vs. [32,10]
	 [[Node: metrics/acc/Equal = Equal[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:GPU:0"](metrics/acc/Reshape, metrics/acc/Cast)]]
	 [[Node: loss/mul/_277 = _Recv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_5751_loss/mul", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:CPU:0"]()]]

[asmekal]

Interesting. Seems it is caused by bug in the latest keras (2.2.3 and 2.2.4) in tensorflow backend (I reproduced this error) with accuracy metric. See related keras issue keras-team/keras#11348

As a simplest fix you can downgrade keras to 2.1.6 - it should work then for sure

Hope keras team will fix it soon, so it should work in the latest keras as well

[SunYanCN]

OK，it works!

[robrechtme]

Do you have any idea if there is a possible fix without downgrading keras?

[asmekal]

@RobrechtM It seems to be an issue with accuracy metric in new version of keras. You can avoid using accuracy and it should be ok.