def __init__(self):
''' Reads the data from training and validation files generated with
@{data_generation_call.generate_data}.
'''
# loadind valid data
self.data = Data()
valid_prefix = arguments.data_path + 'valid'
self.data.valid_mask= np.load(valid_prefix + '.mask')
self.data.valid_mask= self.data.valid_mask:repeatTensor(1,2)
self.data.valid_targets = np.load(valid_prefix + '.targets')
self.data.valid_inputs = np.load(valid_prefix + '.inputs')
self.valid_data_count = self.data.valid_inputs.shape[0]
assert(self.valid_data_count >= arguments.train_batch_size, 'Validation data count has to be greater than a train batch size!')
self.valid_batch_count = self.valid_data_count / arguments.train_batch_size
# loading train data
train_prefix = arguments.data_path + 'train'
self.data.train_mask = np.load(train_prefix + '.mask')
self.data.train_mask = self.data.train_mask:repeatTensor(1,2)
self.data.train_inputs = np.load(train_prefix + '.inputs')
self.data.train_targets = np.load(train_prefix + '.targets')
self.train_data_count = self.data.train_inputs.shape[0]
assert(self.train_data_count >= arguments.train_batch_size, 'Training data count has to be greater than a train batch size!')
self.train_batch_count = self.train_data_count / arguments.train_batch_size
from collections import defaultdict
import numpy as np
import torch
from models import ConEx, ConExWithNorm
from helper_classes import Data
kg_path = 'KGs/WN18RR'
data_dir = "%s/" % kg_path
model_path = 'PretrainedModels/WN18RR/conex_WN18RR.pt'
d = Data(data_dir=data_dir, reverse=False)
class Reproduce:
def __init__(self):
self.cuda = False
self.batch_size = 128
def get_data_idxs(self, data):
data_idxs = [
(self.entity_idxs[data[i][0]], self.relation_idxs[data[i][1]],
self.entity_idxs[data[i][2]]) for i in range(len(data))
]
return data_idxs
def get_er_vocab(self, data):
er_vocab = defaultdict(list)
for triple in data:
er_vocab[(triple[0], triple[1])].append(triple[2])
return er_vocab
from collections import defaultdict
import numpy as np
import torch
from models import ConEx, ConExWithNorm
from helper_classes import Data
kg_path = 'KGs/UMLS'
data_dir = "%s/" % kg_path
model_path = 'PretrainedModels/UMLS/conex_umls.pt'
d = Data(data_dir=data_dir, reverse=True)
class Reproduce:
def __init__(self):
self.cuda = False
self.batch_size = 128
def get_data_idxs(self, data):
data_idxs = [
(self.entity_idxs[data[i][0]], self.relation_idxs[data[i][1]],
self.entity_idxs[data[i][2]]) for i in range(len(data))
]
return data_idxs
def get_er_vocab(self, data):
er_vocab = defaultdict(list)
for triple in data:
er_vocab[(triple[0], triple[1])].append(triple[2])
return er_vocab