Tutorial 04 - Trainning about openff-evaluator HOT 11 CLOSED

@Yu-Tang-Lin it looks like ForceBalance does not understand "single" for the variable jobtype in the optimize.in file. This should be the default option for jobtype, what happens if you try commenting this out in the input file? The alternative (for a single-point calculation) is to set jobtype to optimize and set maxstep to 0.

from openff-evaluator.

Hi,
Sorry, I am checking "optimize.in", Do you mean that comment out this one? (I found it is the most closed code that related to singe)

(allcap) The job type, defaults to a single-point evaluation of objective function
jobtype optimize

from openff-evaluator.

@Yu-Tang-Lin it looks like ForceBalance does not understand "single" for the variable jobtype in the optimize.in file. This should be the default option for jobtype, what happens if you try commenting this out in the input file? The alternative (for a single-point calculation) is to set jobtype to optimize and set maxstep to 0.

I also try maxstep = 0 in optimize.in, but it will show the same error message

from openff-evaluator.

6

@Yu-Tang-Lin it looks like ForceBalance does not understand "single" for the variable jobtype in the optimize.in file. This should be the default option for jobtype, what happens if you try commenting this out in the input file? The alternative (for a single-point calculation) is to set jobtype to optimize and set maxstep to 0.

May you tell me what's your version for forcebalance in your environment? My forcebalance version is 1.9.2

from openff-evaluator.

I have version 1.9.2 installed in my environment. What does your optimize.in file look like?

from openff-evaluator.

Yes, I also have version1.9.2. The optimize.in is generated from code Tutorial 04.
input_file_path = get_data_filename("tutorials/tutorial04/optimize.in")

Here is my optimize.in which was generated from Tutorial 04.
optimize.zip

from openff-evaluator.

Hi, @jeff231li

I also try to modify in forcebalance/optimizer.py (Toolkit in forcebalance)

xk = self.OptTab[self.jobtype]()
It seems that it will extract the jobtype optimize in the optimize.in
But it doesn't work.

I manually add self.jobtype = 'optimize' in forcebalance/optimizer.py
The error message becomes to this. I guess there is some issue in the toolkit in forcebalance for recognizing the jobtype

File "/home/ylinb/anaconda3/envs/OpenFFCrystal/lib/python3.8/site-packages/forcebalance/optimizer.py", line 325, in Run xk = self.OptTab[self.jobtype]() KeyError: 'optimize'

Here is my optimize.err
optimize.txt

Let me know if you need any other info, we can test it together.

from openff-evaluator.

I'm not sure why it's giving you this error. This python script works for me, run the file with this command python run_evaluator.py | tee console_output.log. The problem could be how you installed your conda environment. Can you send the YAML file of your conda environment: conda env export > environment.yaml.

from openff-evaluator.

from openff-evaluator.

Hi @jeff231li,

I follow #414 in the previous issue and use v0.3.9 to test Turorial. It can work. I guess there is an update between v0.3.9 and master.

Thank you for your help!

from openff-evaluator.

I think this has fallen out of date - please do let us know if you run into new issues.

from openff-evaluator.