chrisdonahue / ddc Goto Github PK

Completing the live demo form gives either Internal server error or 413 Request Entity Too Large. Note the uploads seem to go through, and the files are all significantly less than the 16 MB limit.

(I would run the demo locally but I can't seem to get that working either. Thanks for your work on this!)

Uploade a song, select difficulty, type title and artist, and goes to ISE page every time.

Hello, chrisdonahue,

I downloaded itg & fraxtil dataset and I suceeded to generate json files but,
when I analyze itg dataset by running ./smd_4_analyze.sh, the python script stop by ZeroDivisionError in line 83. (it has no error in analyzing fraxtil dataset)
I wonder where does the error come from, or is it correct result?

One of the dependencies name essentia is not found by python when running the bash script ddc_server.sh. I did the following:

brew install essentia (that didn't work)
pip install --user essentia (that didn't work)
installed essentia from source (that didn't solve the problem)

Do you know what other options that I have in order to get this code work on my High Sierra setup on MacOS?

Thanks!

Hi, thanks for sharing this great project! I'm interested in creating a similar project for generating Pump it Up step charts and was thinking of adapting some of the code in this repo as a starting point. Would it be possible to add a license?

With essentia 2.1.beta4 (because I needed to catch up with the latest ffmpeg,) I am getting this error in infer/ddc_server.py:

Creating Mel analyzers
Traceback (most recent call last):
  File "ddc_server.py", line 309, in <module>
    analyzers = create_analyzers(nhop=441)
  File "/Users/cplug/gits/ddc/infer/extract_feats.py", line 20, in create_analyzers
    sampleRate=fs)
  File "/usr/local/opt/essentia/lib/python2.7/site-packages/essentia/standard.py", line 44, in __init__
    self.configure(**kwargs)
  File "/usr/local/opt/essentia/lib/python2.7/site-packages/essentia/standard.py", line 64, in configure
    self.__configure__(**kwargs)
RuntimeError: Error while configuring MelBands: TriangularBands: the number of spectrum bins is insufficient for the specified number of triangular bands. Use zero padding to increase the number of FFT bins.

The MelBands has been malfunctioning on the certain situation as has been discussed on MTG/essentia#142.
Unfortunately, DDC had such situation and latest essentia won't allow this anymore.

Looks like that for the function create_analyzers in extract_feats.py, you either need to double the values of nffts parameter or reduce the value of mel_nband, but both are creating issue elsewhere.
If I do either of those 'fixes,' I get this error when I try to create a chart from my audio. I'll try to read what's going on myself, but for now, I'll report the issue.

Traceback (most recent call last):
  File "ddc_client.py", line 9, in <module>
    print s.create_chart(artist, title, audio_fp, diffs)
  File "/Users/cplug/.pyenv/versions/2.7.14/lib/python2.7/xmlrpclib.py", line 1243, in __call__
    return self.__send(self.__name, args)
  File "/Users/cplug/.pyenv/versions/2.7.14/lib/python2.7/xmlrpclib.py", line 1602, in __request
    verbose=self.__verbose
  File "/Users/cplug/.pyenv/versions/2.7.14/lib/python2.7/xmlrpclib.py", line 1283, in request
    return self.single_request(host, handler, request_body, verbose)
  File "/Users/cplug/.pyenv/versions/2.7.14/lib/python2.7/xmlrpclib.py", line 1316, in single_request
    return self.parse_response(response)
  File "/Users/cplug/.pyenv/versions/2.7.14/lib/python2.7/xmlrpclib.py", line 1493, in parse_response
    return u.close()
  File "/Users/cplug/.pyenv/versions/2.7.14/lib/python2.7/xmlrpclib.py", line 800, in close
    raise Fault(**self._stack[0])
xmlrpclib.Fault: <Fault 1: "<class '__main__.CreateChartException'>:Unknown chart creation exception">

Thanks, @shyamalschandra.

I am getting the following errors when I try to run the code with python2 and 3 with tensorflow 0.12.1 installed along with essentia and other required packages. Should I fix the bug myself or do you know if something is either wrong with my configuration or your configuration?

Shyamals-iMac:infer testuser$ python2 ./ddc_server.py
Traceback (most recent call last):
  File "./ddc_server.py", line 296, in <module>
    if not os.path.isdir(args.out_dir):
  File "/usr/local/Cellar/python/2.7.14_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/genericpath.py", line 49, in isdir
    st = os.stat(s)
TypeError: coercing to Unicode: need string or buffer, NoneType found
Shyamals-iMac:infer testuser$ python3 ./ddc_server.py
  File "./ddc_server.py", line 139
    print 'Extracting metadata from {}'.format(audio_fp)
                                      ^
SyntaxError: invalid syntax

I am currently retraining the models to evaluate my modification to adopt the latest essentia (related: #7.) I want to know a few points about the training process:

1. How long is the training supposed to take?
   I know it mostly depends on the spec, but for the Step Selection model, it took half a day on Core i7 3770 @ 3.40GHz; The Step Placement model, however, is estimated to take over two weeks on the same CPU! I'm not sure if I'm messing something up.
2. After training, which files in/tmp folder should go to server_aux inside infer directory? I found a lot of files there, but none of their names matches the one in server_aux.

Hi Chris,

Thanks for this great work. I was trying to reproduce the reported result with the pretrained model provided in the infer directory. I used Docker so I suppose there is no inconsistency in the environment. I run the model on Fraxtil test set and saved the placed_times for each chart, then used the predicted times and ground truth times to compute F1-c. But I can only get ~0.587 out of that, which is far from the 0.681 reported in your paper, and even worse than LogReg baseline. Is this model actually trained on ITG dataset, or did I do something wrong?

By the way, the link to download ITG dataset is broken. Could you please point me the way to download the dataset?

Thanks!

Hi,

I realized that some of the steps in the setup guide are slightly off and the guides just ends at training stage. I would like to know that how I can get to the final part that evaluates an extra audio piece to generate a stepchart.

Also, it seems that when I do the following
1. Train a step selection (symbolic) model on a dataset: ./sml_sym_2_train.sh fraxtil``

I also needed to do
./sml_sym_1_chart.sh fraxtil

But I don't know whether this is intended.

Thank you!