Update README.md

README.md

@@ -67,7 +67,7 @@ Neural network architecture is defined in [deconvnetlib/network.py](deconvnetlib
 
 ### Generating the database
 
-Example of shell script used to generate a database (system environment variables `LIB_FOLDER` indicates the path to this library - [Beamlib library](https://gitlab.com/daep-ia/beamlib) must be present on it - and should be adapted for each machine):
+Example of shell script used to generate 3 database (training, validation and testing). On each case, 1 source is randomly placed on the scannning grid (21 x 21). System environment variable `LIB_FOLDER` indicates the path to this library - [Beamlib library](https://gitlab.com/daep-ia/beamlib) must be present on it - and should be adapted for each machine.
 
 ```shell
 #!/bin/bash
@@ -76,9 +76,9 @@ conda activate deconvnet
 
 LIB_FOLDER="${HOME}/dev/deeplearning/deconvnet"
 
-python "${LIB_FOLDER}/generate.py" -n=100 -min=1 -max=1 -workers=9 -filename=testing
-python "${LIB_FOLDER}/generate.py" -n=500 -min=1 -max=1 -workers=9 -filename=validation
 python "${LIB_FOLDER}/generate.py" -n=1000 -min=1 -max=1 -workers=9 -filename=training
+python "${LIB_FOLDER}/generate.py" -n=500 -min=1 -max=1 -workers=9 -filename=validation
+python "${LIB_FOLDER}/generate.py" -n=100 -min=1 -max=1 -workers=9 -filename=testing
 ```
 
 | flag | default | comment |