               
			   At last! StarNet++ v2.1.0 module for PI
					     
	        -> TO BE USED FOR ASTRO IMAGE PROCESSING ONLY <-

     This installation will NOT replace original StarNet in PI.

    
======================================================================
                           INSTALLATION
   
   You will need more or less recent version of PI (PCL API Version 160
   or higher). You can go to PROCESSES=>Modules=>Manage Modules and
   look at API Version of any other module. I currently have 171.

1. Close PI.
2. Copy the files from this folder into PixInsight's 'bin' folder.
   Location should be something like 'C:\Program Files\PixInsight\bin'.
   Make sure that StarNet2-pxm.dll is in the 'bin' folder.
3. Run PI.
4. Go to PROCESSES=>Modules=>Install Modules. Hit 'Search'. StarNet
   should appear in the list. Hit 'Install'.
5. You should see StarNet2 module appear under PROCESSES-><Etc> and in
   PRECESSES-><All Processes>.

======================================================================
                          TROUBLESHOOTING
                             
1. If after you hit 'Search' PI does not find StarNet, that means
   that either StarNet2-pxm.dll is not in the right place or you have
   too old PI version. Make sure that StarNet-pxm.dll is in the PI's
   'bin' folder together with other StarNet files and not in some
   subdirectory.
   
2. If after you install the thing you see error message saying
   something about missing checkpoint file, that means that the
   problem is with either 'Starnet2_weights.pb'  file. 
   Make sure they are in the PI's 'bin'
   and not in some subdirectory. Also you can try restarting PI. If
   the problem persists, then there could be some problems while
   copying the files. Try downloading and copying them again.
   
3. If you do all the installation steps correctly, but PI does not
   ever lists StarNet2 module in the search window, that most likely
   means that your CPU does not support AVX instructions and the module
   will not work. You can try using Intel Software Development Emulator
   to emulate those, but expect terrible performance. Note that it is
   not the problem of the build or whatever. Support of AVX
   instructions is crucial for a good performance!
   
   No AVX - no good performance! To see for sure if your CPU supports
   AVX - use software called CPU-Z.

----------------------------------------------------------------------

Dependencies. The code depends on C API of Tensorflow.
   
    https://www.tensorflow.org/install/lang_c

Huge thanks to Daniil Goncharov for his hello_tf_c_api project:

    https://github.com/Neargye/hello_tf_c_api
    
part of which was used in this project.

----------------------------------------------------------------

Please see LICENCE file for legal information.
By using this software you agree to the terms.