def process(self):
data_list = []
indices_val = list(range(225011, 249456))
dp.get_dataset("ZINC_val", regression=True)
node_labels = pre.get_all_node_labels("ZINC_full", True, True)
targets = pre.read_targets("ZINC_val", list(range(0, 24445)))
node_labels_1 = node_labels[225011:249456]
matrices = pre.get_all_matrices_wl("ZINC_val", list(range(0, 24445)))
targets_1 = targets
for i, m in enumerate(matrices):
edge_index_1 = torch.tensor(matrices[i][0]).t().contiguous()
edge_index_2 = torch.tensor(matrices[i][1]).t().contiguous()
data = Data()
data.edge_index_1 = edge_index_1
data.edge_index_2 = edge_index_2
data.x = torch.from_numpy(np.array(node_labels_1[i])).to(
torch.float)
data.y = data.y = torch.from_numpy(np.array([targets_1[i]
])).to(torch.float)
data_list.append(data)
data, slices = self.collate(data_list)
torch.save((data, slices), self.processed_paths[0])
def process(self):
data_list = []
dp.get_dataset("ZINC_test", regression=True)
# TODO Change this
node_labels = pre.get_all_node_labels("ZINC_full", True, True)
targets = pre.read_targets("ZINC_test", list(range(0, 5000)))
node_labels_1 = node_labels[220011:225011]
matrices = pre.get_all_matrices_wl("ZINC_test", list(range(0, 5000)))
targets_1 = targets
for i, m in enumerate(matrices):
edge_index_1 = torch.tensor(matrices[i][0]).t().contiguous()
edge_index_2 = torch.tensor(matrices[i][1]).t().contiguous()
data = Data()
data.edge_index_1 = edge_index_1
data.edge_index_2 = edge_index_2
# one_hot = np.eye(492)[node_labels[i]]
data.x = torch.from_numpy(np.array(node_labels_1[i])).to(
torch.float)
data.y = data.y = torch.from_numpy(np.array([targets_1[i]
])).to(torch.float)
data_list.append(data)
data, slices = self.collate(data_list)
torch.save((data, slices), self.processed_paths[0])
def process(self):
data_list = []
targets = dp.get_dataset("alchemy_full",
multigregression=True).tolist()
node_labels = pre.get_all_node_labels("alchemy_full", True, True)
matrices = pre.get_all_matrices("alchemy_full", list(range(202579)))
for i, m in enumerate(matrices):
edge_index_1 = torch.tensor(matrices[i][0]).t().contiguous()
edge_index_2 = torch.tensor(matrices[i][1]).t().contiguous()
data = Data()
data.edge_index_1 = edge_index_1
data.edge_index_2 = edge_index_2
one_hot = np.eye(83)[node_labels[i]]
data.x = torch.from_numpy(one_hot).to(torch.float)
data.y = data.y = torch.from_numpy(np.array([targets[i]
])).to(torch.float)
data_list.append(data)
data, slices = self.collate(data_list)
torch.save((data, slices), self.processed_paths[0])
def process(self):
data_list = []
targets = dp.get_dataset("QM9", multigregression=True).tolist()
attributes = pre.get_all_attributes("QM9")
node_labels = pre.get_all_node_labels("QM9", False, False)
matrices = pre.get_all_matrices("QM9", list(range(129433)))
for i, m in enumerate(matrices):
edge_index_1 = torch.tensor(matrices[i][0]).t().contiguous()
edge_index_2 = torch.tensor(matrices[i][1]).t().contiguous()
data = Data()
data.edge_index_1 = edge_index_1
data.edge_index_2 = edge_index_2
one_hot = np.eye(3)[node_labels[i]]
data.x = torch.from_numpy(one_hot).to(torch.float)
# Continuous information.
data.first = torch.from_numpy(np.array(
attributes[i][0])[:, 0:13]).to(torch.float)
data.first_coord = torch.from_numpy(
np.array(attributes[i][0])[:, 13:]).to(torch.float)
data.second = torch.from_numpy(
np.array(attributes[i][1])[:, 0:13]).to(torch.float)
data.second_coord = torch.from_numpy(
np.array(attributes[i][1])[:, 13:]).to(torch.float)
data.dist = torch.norm(data.first_coord - data.second_coord,
p=2,
dim=-1).view(-1, 1)
data.edge_attr = torch.from_numpy(np.array(attributes[i][2])).to(
torch.float)
data.y = torch.from_numpy(np.array([targets[i]])).to(torch.float)
data_list.append(data)
data, slices = self.collate(data_list)
torch.save((data, slices), self.processed_paths[0])