Three reels compiled into a set and played by VST instruments: Banjo Accordion
This time we used a database of Irish Traditional Jigs for deep learning music composition to create “neural” ones. All tunes were transposed to the same scale and then feeded to the neural network (Irish-RNN). Learning parameters were used pretty much the same as for the reels generation, so please refer to our previous posts. […]
This time three scales (Edor, Gmaj, Amix) were used to produce a set of Irish Traditional Reels. An increase of a “drop out” parameter during the Irish-RNN training was not really successful so we are rolling back to previous parameters and trying to increase Irish-RNN size. The result is published in ABC and Video format […]
Now it is time for reels! New training parameters are resulted in a big improvement of generated tunes quality. I transposed all “the session” library of reels to the key of D before the network learning. After the training and the generation, for sure, neural Irish tunes require some adjustments of Irish-RNN tunes in ABC […]
The optimization of a learning is a key. To improve the size of a tunes database I transposed all available tunes of a certain types (i.e. reels, jigs, hornpipes) into one scale considering modes (Dorian, Maj and so on). The scale of D was chosen just arbitrary. The very first attempt of the learning process […]
Here I am going to describe one example of a produciton and edition of a tune. The original version was generated by a recurrent neural network trained on Irish traditional tunes in the ABC format. Scpecifically, a small database (~3000 tunes) that consists of irish reels was taken. Let`s take a look at the […]
News from deep learning world. Nvidia presents high quality image translation results on various challenging unsupervised image translation tasks, including street scene image translation, animal image translation, and face image translation. More on the Nvidia website.
Youtube channel of this project is here. All played versions of tunes will be posted on this channel (audio or video).