ValueError: need at least one array to concatenate

[javierrodenas]

Hi!!!
I have the following error training the model :

File ".\ubm.py", line 202, in 
ubm.train()
File ".\ubm.py", line 50, in train
iterations=(1, 2, 2, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8)
File "C:\Users\jrodenas\Desktop\SpeakerRecognition\Speaker-Recognition\sidekit\mixture.py", line 672, in EM_split
self._init(features_server, feature_list, num_thread)
File "C:\Users\jrodenas\Desktop\SpeakerRecognition\Speaker-Recognition\sidekit\mixture.py", line 629, in _init
features = features_server.stack_features_parallel(feature_list, num_thread=num_thread)
File "C:\Users\jrodenas\Desktop\SpeakerRecognition\Speaker-Recognition\sidekit\features_server.py", line 666, in stack_features_parallel
return numpy.concatenate(output, axis=0)
ValueError: need at least one array to concatenate

I have done the data_init first and it was well created. Then I ran extrac_feature and finally, ubm train.
How can I solve that?
Thank you in advance!!

[yangxiaokang]

mybe, you can try it on linux

[Anwarvic]

First, I believe that the features_server is empty and didn't load any features. Why is that? The most probable cause is the location containing the features. To make sure everything as expected, do the following:

• First, open the configuration file conf.yaml and tell me the values of these YAML objects outpath and sampling_rate.
• Then, go to the outpath and you should find at least the following folders audio, feat, and task.
• Inside {outpath}/feat, you should find two folders at least enroll and test. I need to know how many files inside each.
• Also, I need to see the code calling ubm.EM_split() in ubm.py.
• Finally, I need to know the line responsible for creating the FeatureServer inside ubm.py. I'm expecting something like this server = self.createFeatureServer("enroll")... is it right?

[javierrodenas]

@Anwarvic first of all, thanks for your answer.

Answering your questions:

• As outpath I have ./SpeakerRecognition/Speaker-Recognition/Merged_Arabic_Corpus_of_Isolated_Words/ and as sample_rate I have 44100 (default value).

• On the other hand, I find audio, feat and task folders in outpath but inside feat folder I can only see enroll folder but no test folder. Inside the enroll folder there are 6 files:

• enroll_idmap.h5

• plda_idmap.h5

• test_idmap.h5

• test_ndx.h5

• test_trials.txt

• tv_idmap.h5

• Also, ubm.EM_split() has the deault structure:

        ubm.EM_split(
            features_server=server, #sidekit.FeaturesServer used to load data
            feature_list=train_list, #list of feature files to train the model
            distrib_nb=self.NUM_GAUSSIANS, #number of Gaussian distributions
            num_thread=self.NUM_THREADS, # number of parallel processes
            save_partial=False, # if False, it only saves the last model
            iterations=(1, 2, 2, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8)
            )

if __name__ == "__main__":
    conf_filename = "conf.yaml"
    ubm = UBM(conf_filename)
    ubm.train()
    ubm.evaluate()
    ubm.plotDETcurve()
    print( "Accuracy: {}%".format(ubm.getAccuracy()) )

• Finally, you are right. The creation of the FeatureServer is done with server = self.createFeatureServer("enroll")

On the other hand, inside the folder of audio I can find data, enroll and test folders but are empty. Beside this, task folder has the same **6 files ** as feat/enroll.

Thank you in advance.

[Anwarvic]

Now, the problem is that you haven't extracted the features from the data yet. So, follow these steps:

• First, download the data from here.
• Then, run data_init.py. After running it, you will find two folders has been created at {outpath}. these two files are:
  • {outpath}\audio: which will contain two folders at least... enroll and test. Inside each folder you will find audio files that you can listen to.
  • {outpath}\task: which will contain these five files that you have mentioned above.
• Then, you need to run extract_features.py script which will create another directory in the {outpath} called feat. Inside this folder you should find two other folders at least. They are enroll and test.
• After running these two scripts. You can now run ubm.py with no problem.

If you need more information, please check this README.md file as I explained as many details as I could.

[TeppieC]

Hi, maybe it's too late but I believe that you forgot to install sox so the convert_wav() did not work as expected.