Making the music real In development phase, once completed repo will change name to polymuse
Need to discuss .......
Table of Contents
- Installing ..
- Training(Note model)
This is BE project aiming to generate the musical patterns from the midi file that are the audibes to ears
Will be added soon
This to ...
This is pre complete installation, package may not run as expected
$ pip install polymuse
$ pip install polymuse-future
install the polymuse-future recommended
Only NOTE training available
from polymuse import train F = dataset_path # It should be absolute PATH(recomended) where midi file are train.train_gpu(F, maxx = 5) #Only if GPU is available, It uses CuDNNLstm version which performs operation on GPU train.train(F, maxx = 5) #if GPU version do not works
@dataset_path : It should be absolute PATH(recomended) where midi file are @maxx : It is parameters that specifies maximum no of files used to training in case there are large no of files in dataset_path given
This snapshot will train the model on dataset given,
3 files will generated and stored in following dir strucure :
Load Pretrain Models
Below code snapshot downloads the default model, and make above directory structure in current working directory
from polymuse import loader loader.load(mname = 'default')
Load sample midis
Below code snapshot downloads the default midi and download in current directory
from polymuse import loader loader.load_midi()
Before using the player please train the models on dataset or load pre trained models
from polymuse import player # Before this please make sure the h5_models are loaded locally midi_file = "F:\\rushikesh\\project\\dataset\\lakh_dataset\\Kenny G" # directory where midi file will midi_file = dutils.get_all_files(F) # Midi file must be of atleast 3 tracks player.play_3_track_no_time(midi_file, midi_fname = 'midi00')
The above will store midi file in current directory with file name midi00XXX
Release history Release notifications | RSS feed
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Hashes for polymuse_future-0.0.81-py3-none-any.whl