File size: 1,248 Bytes
8346294 4636dfc 8346294 b23da08 592e727 b23da08 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 |
---
license: bsd-3-clause
widget:
- text: "(I%I.M)(piano__0=s0,piano__1=s2,piano__3=s4)"
example_title: "Continue a C-major chord"
---
Musiclang LLM
--------------
Check https://github.com/MusicLang/musiclang to use this model in your musical compositions
MusicLang LLM is a language model based on top of the musiclang symbolic tonal music language. Is is used to predict new music from an existing score.
How to use
----------
Install musiclang package in python :
```
pip install musiclang
```
Predict some music from a given prompt written with musiclang :
```python
from musiclang.library import *
from musiclang import Score
# Some random bar of chopin op18 Waltz
score = ((V % III.b.M)(
piano__0=s0 + s2.e.mp + s3.e.mp,
piano__4=s0.e.o(-2).p + r.e + s0.ed.o(-1).mp + r.s,
piano__5=r + s4.ed.o(-1).mp + r.s,
piano__6=r + s6.ed.o(-1).mp + r.s)+
(V['7'] % III.b.M)(
piano__0=s2.ed.mp + r.s,
piano__2=s4.ed.mp + r.s,
piano__4=s6.ed.o(-1).mp + r.s,
piano__5=s0.ed.o(-1).mp + r.s,
piano__6=s4.ed.o(-1).mp + r.s))
# Predict the next two chords of the score using huggingface musiclang model
predicted_score = score.predict_score(n_chords=2, temperature=0.5)
# Save it to midi
predicted_score.to_midi('test.mid')
```
|