[ase-users] Restarting NEB job with different n_images

[Michael Joseph Waters]

Hi Reza,

You can address the neb images like this:


neb = NEB(images)

neb.images[neb_images//2+1] = saddle.copy()

Correct me if I am wrong, NEB.interpolate doesn't respect existing images other than the first and last? It looks like idpp_interpolate doesn't respect middle images either. But, the SingleCalculatorNEB class has a refinement function.

I wasn't happy with these options for restarting NEBs so a while ago I made my own tool.

Best,
-Mike

________________________________
From: ase-users-bounces at listserv.fysik.dtu.dk <ase-users-bounces at listserv.fysik.dtu.dk> on behalf of Reza Behjatmanesh-Ardakani via ase-users <ase-users at listserv.fysik.dtu.dk>
Sent: Friday, February 21, 2020 2:21 AM
To: ase-users <ase-users at listserv.fysik.dtu.dk>
Subject: [ase-users] Restarting NEB job with different n_images

Hi, I have run a NEB job with a normal image number. However, it seems that there are more steps, and need to increase image numbers. I have saved trajectory file from previous killed job. Now, I want to restart NEB with higher images. How can I do that?
At first I read total images from path.traj file:
images=read('path.traj at -7:')
Now, I need to interpolate between images (increasing 5 to 10 with initial and final becomes 12). I know that I should use NEB.idpp_interpolate(), but don't know how. Unfortunately, there is not example in the site.






-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20200222/a825a379/attachment.html>