A library to generate synthetic tabular or RDF data using Conditional Generative Adversary Networks (GANs) combined with Differential Privacy techniques.
Hello, I read your paper on DP_CGAN and installed DP_CGAN on pytorch. But I got an error when I ran the code example you gave me. I would like to know what solutions you have?
My code:
from dp_cgans import DP_CGAN
import pandas as pd
Hello, I read your paper on DP_CGAN and installed DP_CGAN on pytorch. But I got an error when I ran the code example you gave me. I would like to know what solutions you have? My code: from dp_cgans import DP_CGAN import pandas as pd
tabular_data=pd.read_csv("C:/Users/lenovo/Desktop/workbook1.csv")
model = DP_CGAN(epochs=100)
print("Start training model") model.fit(tabular_data)
Generate 100 synthetic rows
syn_data = model.sample(100) syn_data.to_csv("C:/Users/lenovo/Desktop/workbook.csv")
Error content:
TypeError Traceback (most recent call last) Cell In[4], line 10 7 model = DP_CGAN(epochs=100) 9 print("Start training model") ---> 10 model.fit(tabular_data) 12 # Generate 100 synthetic rows 13 syn_data = model.sample(100)
File E:\anaconda3\envs\pytorch\lib\site-packages\dp_cgans\base.py:143, in BaseTabularModel.fit(self, data) 140 if self._metadata.getdtypes(ids=False): 141 LOGGER.debug( 142 'Fitting %s model to table %s', self.class.name, self.metadata.name) --> 143 self._fit(transformed)
File E:\anaconda3\envs\pytorch\lib\site-packages\dp_cgans\dp_cgan_init.py:59, in DPCGANModel._fit(self, table_data) 55 categories.append(field) 57 # print(table_data[categoricals]) ---> 59 self._model.fit( 60 table_data, 61 discrete_columns=categoricals 62)
File E:\anaconda3\envs\pytorch\lib\site-packages\dp_cgans\synthesizers\dp_cgan.py:488, in DPCGANSynthesizer.fit(self, train_data, discrete_columns, epochs) 485 # condvec = self._data_sampler.sample_condvec(self._batch_size) 487 condvec_pair = self._data_sampler.sample_condvec_pair(self._batch_size) --> 488 c_pair_1, m_pair_1, col_pair_1, opt_pair_1 = condvec_pair 490 # if condvec is None: 491 # c1, m1, col, opt = None, None, None, None 492 # real = self._data_sampler.sample_data(self._batch_size, col, opt) (...) 502 # self._batch_size, col[perm], opt[perm]) 503 # c2 = c1[perm] 505 if condvec_pair is None:
TypeError: cannot unpack non-iterable NoneType object
Looking forward to your reply!