readme

Author: victor-shepardson

Readme.md

@@ -1,8 +1,8 @@
 # Structure
 
 - clients: templates for SuperCollider, Bela (C++), Pure Data, ...
-- osc: 
-- notepredictor: python package for offline pytorch models
+- pytorch-osc: app with OSC server + event loops
+- notepredictor: python package for defining+training pytorch RNN models
 - notebooks: jupyter notebooks
 - scripts: helper scripts for training, data preprocessing etc
 
@@ -11,7 +11,20 @@
 ```
 conda env create -f environment.yml
 conda activate pytorch-osc
-pip install -e pytorch-osc
+pip install -e notepredictor
+```
+
+# Train a model
+
+```
+python scipts/lakh_prep.py --data_path /path/to/midi/files --dest_path /path/to/data/storage
+python scripts/train_pitch.py --data_dir /path/to/data/storage --log_dir /path/for/tensorboard logs --model_dir /path/for/checkpoints train
+```
+
+# Run OSC app
+
+```
+python pytorch-osc/pytorch-osc.py --checkpoint /path/to/my/model.ckpt
 ```
 
 # Develop

clients/pytorch-osc.scd

@@ -37,8 +37,8 @@ o = OSCdef(\test, {
 
 o.free;    // remove the OSCFunc when you are done.
 
-n = NetAddr("127.0.0.1", NetAddr.langPort);
-n.sendMsg('/prediction', 1)
+// n = NetAddr("127.0.0.1", NetAddr.langPort);
+// n.sendMsg('/prediction', 1)
 
 // OSCdef.trace(true)