Add script and instructions

README.md

@@ -17,6 +17,10 @@ A self-supervised denoising algorithm now usable by all in napari.
 Check out the [documentation](https://juglab.github.io/napari-n2v/installation.html) for more detailed installation 
 instructions. 
 
+You can then start the napari plugin by clicking on "Plugins > napari_n2v > Training",
+or run the plugin directly from a [script](scripts/start_plugin.py).
+
+
 
 ## Quick demo
 

docs/installation.md

@@ -17,6 +17,18 @@ operating system.
       pip install "napari[all]==0.4.15" napari-n2v
    ```
 
+> Note: napari-n2v was tested with TensorFlow 2.7 and 2.10, cuda 11.2 and cudnn
+> 8.1.
+
+> **Important**: In order to access the GPU with Tensorflow, it is necessary to
+> export the CUDA library path in your conda environment. Installation 
+> instructions on the TensorFlow website do just that. 
+> For TF 2.10, we recommand running the following in your environment:
+> ```bash
+> mkdir -p $CONDA_PREFIX/etc/conda/activate.d
+> echo 'export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$CONDA_PREFIX/lib/' > $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
+> ```
+
    
 ## macOS
 

scripts/start_plugin.py

@@ -0,0 +1,21 @@
+from pathlib import Path
+import titfffile
+import napari
+
+from napari_n2v import TrainingWidgetWrapper
+
+# create a Viewer
+viewer = napari.Viewer()
+
+# add napari-n2v plugin
+viewer.window.add_dock_widget(TrainingWidgetWrapper(viewer))
+
+# load yout image
+path = Path('path/to/your/image.tif')
+data = titfffile.imread(path)
+
+# add image to napari
+viewer.add_image(data[0][0], name=data[0][1]['name'])
+
+# start UI
+napari.run()