Error when using IS2RE models

[lanchoxie]

Hi, I was using painn_h1024_bs4x8_is2re_all.pt to execute the initial structure to relaxed energy prediction, and I got the following errors: (which did not appear in S2EF models)

from fairchem.core.common.relaxation.ase_utils import OCPCalculator
from ase.io import read
from ase.io import write
import matplotlib.pyplot as plt
from ase.visualize.plot import plot_atoms
from io import StringIO
import contextlib
import os
import numpy as np
import sys

local_model_dir='/home/bingxing2/home/scx8jte/ai4s/fairchem_benchmark/examples/fairchem_checkpoints/'
model='painn_h1024_bs4x8_is2re_all.pt'
id='90614'

system = read(f'vasp_files/combine-{id}.vasp')  

write('initial_structure.vasp', system)
system.set_tags(np.ones(len(system)))  #only when use gemnet

checkpoint_path = os.path.join(local_model_dir,model)  
with contextlib.redirect_stdout(StringIO()) as _:
    calc = OCPCalculator(checkpoint_path=checkpoint_path, cpu=False)
system.set_calculator(calc)
E_slab_adsorbate = system.get_potential_energy()
print(E_slab_adsorbate)

Here is my Error

INFO:root:Loading model: painn
/home/bingxing2/home/scx8jte/.conda/envs/fair-chem/lib/python3.9/site-packages/torch_geometric/nn/conv/message_passing.py:988: UserWarning: 'PaiNNMessage.jittable' is deprecated and a no-op. Please remove its usage.
warnings.warn(f"'{self.class.name}.jittable' is deprecated "
INFO:root:Loaded PaiNN with 72255495 parameters.
INFO:root:Loading checkpoint in inference-only mode, not loading keys associated with trainer state!
INFO:root:Overwriting scaling factors with those loaded from checkpoint. If you're generating predictions with a pretrained checkpoint, this is the correct behavior. To disable this, delete scale_dict from the checkpoint.
/home/bingxing2/home/scx8jte/ai4s/fairchem_benchmark/fairchem/packages/fairchem-core/src/fairchem/core/modules/normalization/normalizer.py:69: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
"mean": torch.tensor(state_dict["mean"]),
WARNING:root:No seed has been set in modelcheckpoint or OCPCalculator! Results may not be reproducible on re-run
/home/bingxing2/home/scx8jte/ai4s/fairchem_benchmark/examples/combined_system_IS2RE_2.py:49: DeprecationWarning: Please use atoms.calc = calc
system.set_calculator(calc)
/home/bingxing2/home/scx8jte/ai4s/fairchem_benchmark/fairchem/packages/fairchem-core/src/fairchem/core/trainers/ocp_trainer.py:461: FutureWarning: torch.cuda.amp.autocast(args...) is deprecated. Please use torch.amp.autocast('cuda', args...) instead.
with torch.cuda.amp.autocast(enabled=self.scaler is not None):
Traceback (most recent call last):
File "/home/bingxing2/home/scx8jte/ai4s/fairchem_benchmark/examples/combined_system_IS2RE_2.py", line 51, in
E_slab_adsorbate = system.get_potential_energy()
File "/home/bingxing2/home/scx8jte/.conda/envs/fair-chem/lib/python3.9/site-packages/ase/atoms.py", line 755, in get_potential_energy
energy = self._calc.get_potential_energy(self)
File "/home/bingxing2/home/scx8jte/.conda/envs/fair-chem/lib/python3.9/site-packages/ase/calculators/abc.py", line 24, in get_potential_energy
return self.get_property(name, atoms)
File "/home/bingxing2/home/scx8jte/.conda/envs/fair-chem/lib/python3.9/site-packages/ase/calculators/calculator.py", line 538, in get_property
self.calculate(atoms, [name], system_changes)
File "/home/bingxing2/home/scx8jte/ai4s/fairchem_benchmark/fairchem/packages/fairchem-core/src/fairchem/core/common/relaxation/ase_utils.py", line 233, in calculate
predictions = self.trainer.predict(batch, per_image=False, disable_tqdm=True)
File "/home/bingxing2/home/scx8jte/.conda/envs/fair-chem/lib/python3.9/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
File "/home/bingxing2/home/scx8jte/ai4s/fairchem_benchmark/fairchem/packages/fairchem-core/src/fairchem/core/trainers/ocp_trainer.py", line 462, in predict
out = self._forward(batch)
File "/home/bingxing2/home/scx8jte/ai4s/fairchem_benchmark/fairchem/packages/fairchem-core/src/fairchem/core/trainers/ocp_trainer.py", line 293, in _forward
raise AttributeError(
AttributeError: Output target: 'forces', not found in model outputs: ['energy']

[misko]

Hi!
Thank you for this, you are totally right, this is an issue! I have created a fix here , #922 , and will hopefully have this merged soon.

In the meantime if this is blocking you please try to use the branch, add_calculator_only_output_field

[github-actions]

This issue has been marked as stale because it has been open for 30 days with no activity.