tensorflow.python.framework.errors_impl.InvalidArgumentError: All dimensions except 2 must match. Input 1 has shape [128 32 512] and doesn't match input 0 with shape [128 16 512]

[xiaojian10]

Hi！
I recently encountered a problem as follows:
List of arrays in input file: ['data', 'label']
Shape of X: (66688, 8192, 1)
Shape of Y: (66688, 8192, 1)
List of arrays in input file: ['data', 'label']
Shape of X: (5248, 8192, 1)
Shape of Y: (5248, 8192, 1)
audiotfilm
building model...
D-Block: (None, None, 128)
D-Block: (None, None, 256)
D-Block: (None, None, 512)
D-Block: (None, None, 512)
U-Block: (None, None, 1024)
U-Block: (None, None, 1024)
U-Block: (None, None, 512)
U-Block: (None, None, 256)
creating train_op with params: {'alg': 'adam', 'lr': 0.001, 'b1': 0.9, 'b2': 0.999, 'batch_size': 128, 'layers': 4}
WARNING:tensorflow:From /root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/util/tf_should_use.py:247: initialize_all_variables (from tensorflow.python.ops.variables) is deprecated and will be removed after 2017-03-02.
Instructions for updating:
Use tf.global_variables_initializer instead.
Parameters: 68221186
Traceback (most recent call last):
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1375, in _do_call
return fn(*args)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1360, in _run_fn
target_list, run_metadata)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1453, in _call_tf_sessionrun
run_metadata)
tensorflow.python.framework.errors_impl.InvalidArgumentError: All dimensions except 2 must match. Input 1 has shape [128 32 512] and doesn't match input 0 with shape [128 16 512].
[[{{node gradients/generator/upsc_conv3/concatenate/concat_grad/ConcatOffset}}]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "run.py", line 176, in
main()
File "run.py", line 173, in main
args.func(args)
File "run.py", line 125, in train
model.fit(X_train, Y_train, X_val, Y_val, n_epoch=args.epochs, r=args.r, speaker=args.speaker, grocery=args.grocery, piano=args.piano, calc_full_snr = full)
File "/root/xiaojian/two/audio-super-res-master/src/models/model.py", line 249, in fit
tr_objective = self.train(feed_dict)
File "/root/xiaojian/two/audio-super-res-master/src/models/model.py", line 324, in train
[self.train_op, self.loss], feed_dict=feed_dict)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 968, in run
run_metadata_ptr)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1191, in _run
feed_dict_tensor, options, run_metadata)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1369, in _do_run
run_metadata)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1394, in _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: All dimensions except 2 must match. Input 1 has shape [128 32 512] and doesn't match input 0 with shape [128 16 512].
[[node gradients/generator/upsc_conv3/concatenate/concat_grad/ConcatOffset (defined at /root/xiaojian/two/audio-super-res-master/src/models/model.py:151) ]]

Is this problem caused by the environment configuration, or is it?

[xiaojian10]

Original stack trace for 'gradients/generator/upsc_conv3/concatenate/concat_grad/ConcatOffset':
File "run.py", line 176, in
main()
File "run.py", line 173, in main
args.func(args)
File "run.py", line 123, in train
model = get_model(args, n_dim, r, from_ckpt=False, train=True)
File "run.py", line 159, in get_model
strides=args.strides, opt_params=opt_params, log_prefix=args.logname)
File "/root/xiaojian/two/audio-super-res-master/src/models/audiotfilm.py", line 31, in init
opt_params=opt_params, log_prefix=log_prefix)
File "/root/xiaojian/two/audio-super-res-master/src/models/model.py", line 55, in init
self.train_op = self.create_train_op(X, Y, alpha)
File "/root/xiaojian/two/audio-super-res-master/src/models/model.py", line 91, in create_train_op
grads = self.create_gradients(self.loss, params)
File "/root/xiaojian/two/audio-super-res-master/src/models/model.py", line 151, in create_gradients
gv = self.optimizer.compute_gradients(loss, params)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/training/optimizer.py", line 512, in compute_gradients
colocate_gradients_with_ops=colocate_gradients_with_ops)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/gradients_impl.py", line 172, in gradients
unconnected_gradients)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/gradients_util.py", line 684, in _GradientsHelper
lambda: grad_fn(op, *out_grads))
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/gradients_util.py", line 340, in _MaybeCompile
return grad_fn() # Exit early
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/gradients_util.py", line 684, in
lambda: grad_fn(op, *out_grads))
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/array_grad.py", line 229, in _ConcatGradV2
op, grad, start_value_index=0, end_value_index=-1, dim_index=-1)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/array_grad.py", line 155, in _ConcatGradHelper
offset = gen_array_ops.concat_offset(non_neg_concat_dim, sizes)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/gen_array_ops.py", line 1140, in concat_offset
"ConcatOffset", concat_dim=concat_dim, shape=shape, name=name)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py", line 750, in _apply_op_helper
attrs=attr_protos, op_def=op_def)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/framework/ops.py", line 3536, in _create_op_internal
op_def=op_def)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/framework/ops.py", line 1990, in init
self._traceback = tf_stack.extract_stack()

...which was originally created as op 'generator/upsc_conv3/concatenate/concat', defined at:
File "run.py", line 176, in
main()
[elided 3 identical lines from previous traceback]
File "/root/xiaojian/two/audio-super-res-master/src/models/audiotfilm.py", line 31, in init
opt_params=opt_params, log_prefix=log_prefix)
File "/root/xiaojian/two/audio-super-res-master/src/models/model.py", line 48, in init
self.predictions = self.create_model(n_dim, r)
File "/root/xiaojian/two/audio-super-res-master/src/models/audiotfilm.py", line 135, in create_model
x = Concatenate()([x, l_in])
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/keras/engine/base_layer_v1.py", line 786, in call
outputs = call_fn(cast_inputs, *args, **kwargs)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/keras/layers/merge.py", line 183, in call
return self._merge_function(inputs)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/keras/layers/merge.py", line 522, in _merge_function
return K.concatenate(inputs, axis=self.axis)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/util/dispatch.py", line 201, in wrapper
return target(*args, **kwargs)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/keras/backend.py", line 2989, in concatenate
return array_ops.concat([to_dense(x) for x in tensors], axis)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/util/dispatch.py", line 201, in wrapper
return target(*args, **kwargs)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/array_ops.py", line 1677, in concat
return gen_array_ops.concat_v2(values=values, axis=axis, name=name)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/ops/gen_array_ops.py", line 1208, in concat_v2
"ConcatV2", values=values, axis=axis, name=name)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py", line 750, in _apply_op_helper
attrs=attr_protos, op_def=op_def)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/framework/ops.py", line 3536, in _create_op_internal
op_def=op_def)
File "/root/anaconda3/envs/xiaojian/lib/python3.7/site-packages/tensorflow/python/framework/ops.py", line 1990, in init
self._traceback = tf_stack.extract_stack()

[Sawyerb]

What's the full command you're using to train? I can't reproduce the issue on my end. What are you setting as the pool size and stride?

[xiaojian10]

Thank you very much for your reply, my parameter settings are as follows：

The .h5 file is generated by executing Makefile, and the detailed parameters are:

Second, in the training model, my input is:

The remaining parameters are the default parameters in run.py！

[xiaojian10]

If the .h5 file obtained by using the default parameter settings in the Makefile, the following problems will occur:

[Sawyerb]

Sorry for the delayed response. Rather than using the default parameters, use the following:

  --train ../data/vctk/speaker1/vctk-speaker1-train.4.16000.8192.4096.h5 \
  --val ../data/vctk/speaker1/vctk-speaker1-val.4.16000.8192.4096.h5 \
  -e 120 \
  --batch-size 64 \
  --lr 3e-4 \
  --logname singlespeaker \
  --model audiotfilm \
  --r 4 \
  --layers 4 \
  --piano false \
  --pool_size 2 \
  --strides 2
  --full true

[Sawyerb]

And you can't train data files with dimension -1. (Those are just used to evaluate the "full" SNR on non-patched audio.)

If you're just looking to test out the code, I recommend sticking with the example code in the README.