def __init__(self, path, cutoff, properties = [], subset=None, collect_triples = False):
self.path = path
self.cutoff = cutoff
self.properties = properties
self.collect_triples = collect_triples
self.dbpath = os.path.join(self.path, 'icsd.db')
environment_provider = ASEEnvironmentProvider(cutoff)
super(ICSD, self).__init__(self.dbpath, subset, properties, environment_provider, False)
def __init__(self, path, cutoff, apikey=None, download=True, subset=None, properties=[], collect_triples=False):
self.path = path
self.cutoff = cutoff
self.apikey = apikey
if not os.path.exists(self.path):
os.makedirs(self.path)
self.dbpath = os.path.join(self.path, 'mp.db')
environment_provider = ASEEnvironmentProvider(cutoff)
if download and not os.path.exists(self.dbpath):
self._download()
super(MaterialsProject, self).__init__(self.dbpath, subset, properties, environment_provider, collect_triples)
def neighbor_gen(at,
distance_expansion=None,
cutoff=5.0,
n_gaussians=25,
trainable_gaussians=False,
environment_provider=ASEEnvironmentProvider(5.0),
collect_triples=False,
pair_provider=None,
center_positions=True):
properties = {}
properties[Structure.Z] = tensor(at.numbers.astype(np.int)).unsqueeze(0)
positions = at.positions.astype(np.float32)
if center_positions:
positions -= at.get_center_of_mass()
properties[Structure.R] = tensor(positions).unsqueeze(0)
properties[Structure.cell] = tensor(at.cell.astype(
np.float32)).unsqueeze(0)
# get atom environment
idx = 0
nbh_idx, offsets = environment_provider.get_environment(idx, at)
properties[Structure.neighbors] = tensor(nbh_idx.astype(
np.int)).unsqueeze(0)
properties[Structure.cell_offset] = tensor(offsets.astype(
np.float32)).unsqueeze(0)
properties[Structure.neighbor_mask] = None
properties['_idx'] = tensor(np.array([idx], dtype=np.int)).unsqueeze(0)
if collect_triples:
nbh_idx_j, nbh_idx_k = collect_atom_triples(nbh_idx)
properties[Structure.neighbor_pairs_j] = tensor(
nbh_idx_j.astype(np.int))
properties[Structure.neighbor_pairs_k] = tensor(
nbh_idx_k.astype(np.int))
model = spk.custom.representation.RBF(
distance_expansion=distance_expansion,
cutoff=cutoff,
n_gaussians=n_gaussians,
trainable_gaussians=trainable_gaussians)
model = to_device(model)
r, f = model.forward(properties)
return to_np(r.squeeze()), to_np(f.squeeze())
def __init__(self,
path,
cutoff,
properties=[],
subset=None,
collect_triples=False):
self.path = path
self.cutoff = cutoff
self.properties = properties
self.collect_triples = collect_atom_triples_offsets
if self.path.rsplit(".", 1)[-1] == "db": self.dbpath = self.path
else: self.dbpath = os.path.join(self.path, 'perovskites.db')
environment_provider = ASEEnvironmentProvider(cutoff)
super(Perovskites,
self).__init__(self.dbpath, subset, properties,
environment_provider, collect_triples)
# ax[0,2].set_xlabel('Relative error (eV/atom)',weight='bold',fontsize=14)
# ax[1,0].hist(forces,bins=fbins,label='DFT')
# ax[1,0].hist(forceresults,bins=fbins, color='orange',alpha=0.7,label='SchNet')
# ax[1,0].legend()
# ax[1,0].set_xlabel('Forces (eV/$\AA$)',weight='bold',fontsize=14)
# ax[1,1].hist(fdiff,bins=fbins)
# ax[1,1].set_xlabel('Error (eV/$\AA$)',weight='bold',fontsize=14)
# ax[1,2].hist(frelative,bins=fbins)
# ax[1,2].set_xlabel('Relative error (eV/$\AA$)',weight='bold',fontsize=14)
# fig.suptitle('Error report for ' + file,weight='bold',fontsize=18)
# plt.show()
device = torch.device("cuda")
train_args = args
spk.utils.set_random_seed(args.seed)
env = ASEEnvironmentProvider(args.cutoff)
#name = r'/dev/shm/data/train' + sys.argv[1]
name = r'/dev/shm/data/bulkVtrain3200'
data_train = schnetpack2.custom.datasets.extxyz1.ExtXYZ(
name + '.db',
name + '.xyz',
environment_provider=env,
properties=['energy', 'forces'])
#name = r'/dev/shm/data/test' + sys.argv[1]
name = r'/dev/shm/data/bulkVtest3200'
data_val = schnetpack2.custom.datasets.extxyz1.ExtXYZ(
name + '.db',
name + '.xyz',
environment_provider=env,
properties=['energy', 'forces'])
arr_temp = np.ctypeslib.as_array(output_ptr, shape=arr_shape)
return arr_temp[:, :,
0], arr_temp[:, :,
1:] #[output_ptr[i] for i in range(N.value*max_neighbours.value*4)]#np.ctypeslib.as_array(output_ptr, shape = arr_shape) #make_nd_array(output_ptr, arr_shape)
batch_size = 16
cutoff = 5.0
apikey = None
datapath = "/scratch/leuven/412/vsc41276/mp.db"
property = "formation_energy_per_atom"
mp_ASE = AtomsData(datapath,
properties=[property],
environment_provider=ASEEnvironmentProvider(cutoff))
mp_MAARTEN = AtomsData(datapath,
properties=[property],
environment_provider=MaartenEnvironmentProvider(cutoff))
ASE_loader = spk.data.AtomsLoader(mp_ASE,
batch_size=batch_size,
sampler=RandomSampler(mp_ASE),
num_workers=36,
pin_memory=True)
MAARTEN_loader = spk.data.AtomsLoader(mp_MAARTEN,
batch_size=batch_size,
sampler=RandomSampler(mp_ASE),
num_workers=36,