StarNet - new module for star removal

mguandalini said:
Hi I'm Marco
I have two PC, a latop and desktop
When i install starnet on desktop, i have no problem, PI find a new module without error
On my latop Olivetti i try more time to search a module but PI don't recognise the form, even if the files are copied to the right folder
Please can you help me to solve this problem?
Thanks
Marco
Click to expand...

Hi, Marco!

That means that the code will not work on your laptop! You can find more details in my posts above in this thread and in the README file distributed with StarNet to get more info on why!
 
  I had great hopes that this module would work.  I tried it on two different computers where I loaded into program files for PI bin and no luck. All files were present as shown in the readme section of the download.  Must be the cpu limitation.  Is there anyway that this can be further refined so that it operates with all cpu's?
 
gunny said:
  I had great hopes that this module would work.  I tried it on two different computers where I loaded into program files for PI bin and no luck. All files were present as shown in the readme section of the download.  Must be the cpu limitation.  Is there anyway that this can be further refined so that it operates with all cpu's?
Click to expand...

This "limitation" isn't there for no reason. It's crucial to get decent performance and is very unlikely to go away. Please read README file for more information!
 
not yet... the only module that exists right now is for windows. however there is a standalone program called starnet++ that runs on macosx. you need some familiarity with the osx command line to use it though.

rob
 
pfile said:
not yet... the only module that exists right now is for windows. however there is a standalone program called starnet++ that runs on macosx. you need some familiarity with the osx command line to use it though.

rob
Click to expand...

And I can confirm that starnet++ on MacOS works perfectly well

Dave
 
I am trying to run StarNet++ on my MAC OS but have never used the terminal before.  I have manage to open the Terminal and change the directory and list the contents of the folder I was assumming the something like exec run_rgb_starnet.sh would run the file but it didn't work.  Could you help me?
 
you don't need exec, you can either do:

chmod +x ./run_rgb_starnet.sh
./run_rgb_starnet.sh

that's going to run starnet with canned input and output filenames and a stride of 64. the input filename will have to be "rgb_test5.tif" (no quotes of course) and it will output the starless version in "rgb_test5_s.tif". you only have to do the chmod thing once.

if you want to just run starnet on some arbitrary file:

chmod +x ./rgb_starnet++
./rgb_starnet++ your_input_filename.tif whatever_output_filename_you_like.tif

i think nikita has recommended to just run with the default stride, so i have not put it on the command line shown above. also same as the first command, you only need to do the chmod +x thing one time.

rob
 
Thank you Rob,

This the third or fouth time you've helped me over the years and I appreciate it!

I successfully ran the Mono script and got the starless output.  I then tried to repeat the process.  It didn't run.  See below:

[size=14pt]RobertMac:StarNet_MacOS robertkimball$ chmod +x ./rgb_starnet++
RobertMac:StarNet_MacOS robertkimball$ ./run_rgb_starnet.sh
./run_rgb_starnet.sh: line 4: ./rgb_starnet++: cannot execute binary file
RobertMac:StarNet_MacOS robertkimball$
[/size]
 
no problem on the help!

looks like you still need to do chmod +x ./run_rgb_starnet.sh

but given where you are, i think you should just skip the shell script. thinking about it more, i think the script was included in the zipfile so that one could double-click the .sh from the finder and it would just run. but that requires you to rename your input images to rgb_test5.tif every time you run it. since you're already successfully in Terminal.app and in the StarNet_MacOS directory, i think you might as well just run

./rgb_starnet++ your_input_image_name.tif whatever_you_want_to_call_the_output.tif

so maybe something like like

./rgb_starnet++ tadpoles_rgb.tif tadpoles_rgb_starless.tif

rob

 
It's interesting that the RGB_starnet++ is of type "Alias" and the mono_starnet++ is of type "Unix Executible."  Could that be the problem?  I've got the mono_starnet working perfectly.
 
yeah, both of course should be executables... not sure how yours became a link?

pfile% file mono_starnet++ rgb_starnet++
mono_starnet++: Mach-O 64-bit x86_64 executable, flags:<NOUNDEFS|DYLDLINK|TWOLEVEL|WEAK_DEFINES|BINDS_TO_WEAK|PIE>
rgb_starnet++:  Mach-O 64-bit x86_64 executable, flags:<NOUNDEFS|DYLDLINK|TWOLEVEL|WEAK_DEFINES|BINDS_TO_WEAK|PIE>

if you do

ls -l rgb_starnet++

you will see what it is pointing to.

rob


 
It's the stangest thing!  I unzipped the folder again and the rgb_starnet++ was listed as Unix Executable.  It is running right now in the background.  Thanks so much for all your help.
 
Having big fun now that Rob has helped me get StarNet++ running.  I have attached two examples.  They look much better when they are seen as 16 bit tifs.
 

Attachments

  • Iris_NS.jpg
    Iris_NS.jpg
    222.9 KB · Views: 104
  • Eagle_NS.jpg
    Eagle_NS.jpg
    198.1 KB · Views: 89
Downloaded the module and moved all files into the Pixinsight bin directory. No new modules available, tried search, same thing. I'm running a pretty new workstation, Xeon processor and Quadro P5000 16 Gig graphics card. I'm new to module install, not sure what I'm missing
 
i wonder if the Xeon processors don't have all the vector instructions that the starnet module needs. a sanity check is to download starnet++ and see if it runs.

https://sourceforge.net/projects/starnet/

rob
 
Thanks! I did just run it from the .bat file included in the download and it worked fine. It just said the version of tensorflow included isn?t compiled to use the AVX2 extensions of the processor, but then went ahead and worked. All you are supposed to do is copy all the files in the download to the PI .bin directory then open the PI Process-> Modules command, right?

Wsyne
 
ok so that warning just means that if you come up with a tensorflow.dll that better supports your cpu it will run faster. i think the PI module checks if it can run and won?t install if it can?t, the most common reason being that the cpu lacks the minimum set of vector instructions. so that?s not the problem you?re having.

as for how to install the PI module, i am not sure since i usually use macosx and haven?t tried installing it on my windows machine (the module only exists for windows at this time.)

rob
 
You must log in or register to reply here.
Back
Top