cifkao / groove2groove Goto Github PK

Hi. Thank you for sharing your code.
I run your code as you said and use the below command, replacing the $LOGDIR with the model directory.
However, I got this error

anaconda3/envs/groove2groove/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 1534, in restore
raise ValueError("Can't load save_path when it is None.")
ValueError: Can't load save_path when it is None.

Replace $LOGDIR with the model directory, containing the model.yaml configuration file (e.g. one of the directories under experiments).

To run the trained model on a single pair of MIDI files, use the run-midi command, e.g.:

python -m groove2groove.models.roll2seq_style_transfer --logdir $LOGDIR run-midi
--sample --softmax-temperature 0.6
content.mid style.mid output.mid

Hi. Thank you for sharing your code.
I run your code as you said and use the below command, replacing the $LOGDIR with the model directory.
However, I got this error

Replace $LOGDIR with the model directory, containing the model.yaml configuration file (e.g. one of the directories under experiments).

To run the trained model on a single pair of MIDI files, use the run-midi command, e.g.:

python -m groove2groove.models.roll2seq_style_transfer --logdir $LOGDIR run-midi --sample home/project/Desktop/Groove2groove/groove2groove/experiments/v01_drums_vel/v01_drums_vel/latest_checkpoint --softmax-temperature 0.6 content.mid style.mid output.mid

When i run this command, the following error occured.
OSError: no MTrk header at start of track

I downloaded the data set you shared via this website : [https://zenodo.org/record/3958000#.YNG5YXUzaV4]

I found this midi file like "maj500.BRITROK8_b.midi"

I guess maj500 is content and BRITROK8 is a style file.
And this midi file is an output of style embedding. Is that correct?

But not sure why there is a substyle a or b.
Could you explain this?
and where can I download the style file BRITROK8?

Thank you

I think training is succeded somehow. There was no error during training.
And.. I just tested it with the midi files you showed on the demo web page.
Two input (content input, style input)

However, I get the below two errors. I think one is from rnn layers and not sure for the other one.

2021-04-23 22:51:39 seonghyeon-cbim tensorflow[53024] INFO Restoring parameters from experiments/v01/latest.ckpt-9600
2021-04-23 22:51:39 seonghyeon-cbim museflow[53024] INFO Variables restored from experiments/v01/latest.ckpt-9600
16it [00:00, 33.97it/s]
Traceback (most recent call last):
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1334, in _do_call
return fn(*args)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1319, in _run_fn
options, feed_dict, fetch_list, target_list, run_metadata)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1407, in _call_tf_sessionrun
run_metadata)
tensorflow.python.framework.errors_impl.UnimplementedError: TensorArray has size zero, but element shape [?,200] is not fully defined. Currently only static shapes are supported when packing zero-size TensorArrays.
[[{{node encoder_rnn/rnn/TensorArrayStack/TensorArrayGatherV3}} = TensorArrayGatherV3[dtype=DT_FLOAT, element_shape=[?,200], _device="/job:localhost/replica:0/task:0/device:CPU:0"](encoder_rnn/rnn/TensorArray, encoder_rnn/rnn/TensorArrayStack/range, encoder_rnn/rnn/while/Exit_2)]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/runpy.py", line 193, in _run_module_as_main
"main", mod_spec)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/groove2groove/models/roll2seq_style_transfer.py", line 407, in
main()
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/groove2groove/models/roll2seq_style_transfer.py", line 401, in main
args.func(experiment, args)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/groove2groove/models/roll2seq_style_transfer.py", line 185, in run_midi
sequences = self._run_cli(args, pipeline)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/groove2groove/models/roll2seq_style_transfer.py", line 197, in _run_cli
sample=args.sample, softmax_temperature=args.softmax_temperature)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/groove2groove/models/roll2seq_style_transfer.py", line 211, in run
self.trainer.session, dataset, sample, softmax_temperature, options=options) or []
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/groove2groove/models/roll2seq_style_transfer.py", line 122, in run
options=options)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/museflow/model_utils.py", line 170, in run_over_dataset
options=options))
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/museflow/model_utils.py", line 139, in run
return session.run(fetches, feed_dict, options=options)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 929, in run
run_metadata_ptr)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1152, in _run
feed_dict_tensor, options, run_metadata)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1328, in _do_run
run_metadata)
File "/home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1348, in _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.UnimplementedError: TensorArray has size zero, but element shape [?,200] is not fully defined. Currently only static shapes are supported when packing zero-size TensorArrays.
[[node encoder_rnn/rnn/TensorArrayStack/TensorArrayGatherV3 (defined at /home/seonghyeon/anaconda3/envs/groove2groove/lib/python3.6/site-packages/museflow/components/rnn_layer.py:47) = TensorArrayGatherV3[dtype=DT_FLOAT, element_shape=[?,200], _device="/job:localhost/replica:0/task:0/device:CPU:0"](encoder_rnn/rnn/TensorArray, encoder_rnn/rnn/TensorArrayStack/range, encoder_rnn/rnn/while/Exit_2)]]

I need an output that should maintain some of the content input style.
So I can tell it's from the content.
However, after style transfer, it's difficult to catch the content input style from the output.

Is there a way to conserve input music like 40 ~ 50 percent and add a different style?

Maybe this is the limitation of the current model?
I found that you added content input to an output midi again at the last stage in the demo site.

I have download Band-in-a-box and refer to https://github.com/cifkao/pybiab. But there are still some problem that I can't fix. For example, I find the generation result from Real-Band is a little different every time. There are different notes every time I save the same song with the same style. And how to fix the wrong program number in MIDI files?
Thank you very much!

I'm getting this message always when I try to use the demo. What's happening?

I assume you've got the best trained model from the configure in experiment/v01_drum.
I wonder how long I should train this model to reach your model on the demo website.

My best model is at checkpoint 8000 ~ 9000 and the loss is around 130~140.

Should I train more? or can you provide a pre-trained model?