Comments (7)
Hi! Sorry, I am not sure I understand what you changed?
could you perhaps explain a little more?
from mace.
Actually I think there might be a mistake in the function I wrote.
It works fine for training, but in the MD prediction it gives an error.
The hole idea of my implementation was to be able to use ASE trajectory files :) Do you think that can be made possible?
from mace.
Could you post the error you get when trying to load your traj file? If ase.io.read works the same way on them, I am not sure where the problem might be.
from mace.
The energy and forces will not be loaded, since they do not have the atoms.info and atoms.arrays dictionary :) The will not be an error by it self but the model is not training.
from mace.
The problem with the command extraction is that it can be unreliable. If a calculator is attached to the atom object, it can give inconsistent results. For this reason, storing your data in the info with a custom key is highly recommended. In particular, .traj files are usually associated with the calculator that generated them. Is there an easy way to convert the .traj to .xyz? We could add this pre-processing.
from mace.
I think traj files store those quantities in a SinglePointCalculator
, which is an essentially static storage that fills in the return values for get_potential_energy()
etc. In fact, it's probably safer than quantities from extxyz, because I think the calculator is invalidated if you modify positions etc. It's just not possible to do things like save multiple sets of results from different calculators. It would, of course, be trivial to write a little loop that reads the calculator and stores the results in info/arrays. The ase.io extxyz reader probably already does that, I think.
from mace.
I think converting traj to xyz file will require mor space since you will have the same data file twice.
What Bernstei suggest would be a great solution!
from mace.
Related Issues (20)
- Restart of the training using --restart_latest doesn't properly work - resolved -
- Deployment of the MACE Model Without Using ASE HOT 5
- Error while using multihead interface HOT 2
- Segmentation fault during LAMMPS-MACE-CPU run HOT 1
- Problems with parallelization on CPU (Using LAMMPS) HOT 1
- Improve plotting script
- Automatically add parameters to optimizer
- Change stress to virials internally
- ASE documentation update? HOT 4
- Move to ruff for linting and formating? HOT 3
- Tripeptides Data for MACE-OFF Model
- Bug on loading finetuning model HOT 2
- Remove `e3nn` version pin HOT 2
- models created from multihead fine tuning don't work in lammps HOT 6
- universal loss (at least in multi-head-interface branch) does not support per-config weight
- cannot turn off multihead finetuning in multi-head-interface branch HOT 17
- Colab Tutorial Link Does not work HOT 2
- help message for `--pair_repulsion` is wrong HOT 1
- How to use multi-GPUs training with PBS system HOT 20
- [feature suggest] active learning in lammps
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from mace.