I got this error while trying to learn tabddpm

[limhasic]

from synthcity.plugins import Plugins
from synthcity.plugins.core.dataloader import GenericDataLoader
from synthcity.utils.serialization import save_to_file, load_from_file
from sklearn.ensemble import VotingClassifier
from sklearn.metrics import f1_score, accuracy_score
from tqdm import tqdm
import pandas as pd
import numpy as np
import optuna
from optuna.samplers import TPESampler
from optuna.pruners import SuccessiveHalvingPruner
from sklearn.metrics import f1_score
from lightgbm import LGBMClassifier
import json
import os
os.environ['CUDA_LAUNCH_BLOCKING'] = "1"
os.environ["CUDA_VISIBLE_DEVICES"] = "0"

from synthcity.utils.optuna_sample import suggest_all
tabddpm_model = Plugins().get("ddpm")
trial = optuna.create_study().ask()
params = suggest_all(trial, tabddpm_model.hyperparameter_space())

params['lr'] = 0.0001
params['num_timesteps'] = 100 # 데이터 생성 과정에서 30단계의 시간을 통해 모델이 데이터를 생성
params['batch_size'] = 10
params['n_iter'] = 1000

train_data = pd.read_csv('train_12000.csv').drop('Unnamed: 0', axis=1)

subset -> train_data

loader = GenericDataLoader(train_data, target_column="Fraud_Type")

TabDDPM 모델 초기화 및 학습

PLUGIN = "ddpm"
plugin_cls = type(Plugins().get(PLUGIN))
plugin = plugin_cls(**params).fit(loader)
#save_to_file(f'ddpm_{fraud_type}.pkl', syn_model)
#save_to_file(f'ddpm_{fraud_type}.pkl', plugin)
save_to_file(f'ddpm_all.pkl', plugin)
#reloaded = load_from_file('./adsgan_10_epochs.pkl')

합성 데이터 생성

#synthetic_subset = plugin.generate(count=N_Syn_len)
print('done')

I got this error while trying to learn tabddpm

Epoch: 0%| | 0/1000 [00:00<?, ?it/s]../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [0,0,0], thread: [0,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [0,0,0], thread: [1,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [0,0,0], thread: [3,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [0,0,0], thread: [7,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [0,0,0], thread: [8,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
Epoch: 0%| | 0/1000 [00:00<?, ?it/s]

RuntimeError Traceback (most recent call last)
Cell In[1], line 38
36 PLUGIN = "ddpm"
37 plugin_cls = type(Plugins().get(PLUGIN))
---> 38 plugin = plugin_cls(**params).fit(loader)
39 #save_to_file(f'ddpm_{fraud_type}.pkl', syn_model)
40 #save_to_file(f'ddpm_{fraud_type}.pkl', plugin)
41 save_to_file(f'ddpm_all.pkl', plugin)

File /opt/venv/lib/python3.10/site-packages/pydantic/decorator.py:40, in pydantic.decorator.validate_arguments.validate.wrapper_function()

File /opt/venv/lib/python3.10/site-packages/pydantic/decorator.py:134, in pydantic.decorator.ValidatedFunction.call()

File /opt/venv/lib/python3.10/site-packages/pydantic/decorator.py:206, in pydantic.decorator.ValidatedFunction.execute()

File /opt/venv/lib/python3.10/site-packages/synthcity/plugins/core/plugin.py:256, in Plugin.fit(self, X, *args, **kwargs)
248 X, self.compress_context = load_from_file(bkp_file)
250 self._training_schema = Schema(
251 data=X,
252 sampling_strategy=self.sampling_strategy,
253 random_state=self.random_state,
254 )
--> 256 output = self._fit(X, *args, **kwargs)
257 self.fitted = True
259 return output

File /opt/venv/lib/python3.10/site-packages/synthcity/plugins/generic/plugin_ddpm.py:230, in TabDDPMPlugin._fit(self, X, *args, **kwargs)
227 self.expecting_conditional = True
229 # NOTE: cond may also be included in the dataframe
--> 230 self.model.fit(df, cond, **kwargs)
231 self.loss_history = self.model.loss_history
232 self.validation_history = self.model.val_history

File /opt/venv/lib/python3.10/site-packages/synthcity/plugins/core/models/tabular_ddpm/init.py:163, in TabDDPM.fit(self, X, cond, **kwargs)
161 self.optimizer.zero_grad()
162 args = (x,) if cond is None else (x, y)
--> 163 loss_multi, loss_gauss = self.diffusion.mixed_loss(*args)
164 loss = loss_multi + loss_gauss
165 loss.backward()

File /opt/venv/lib/python3.10/site-packages/synthcity/plugins/core/models/tabular_ddpm/gaussian_multinomial_diffsuion.py:665, in GaussianMultinomialDiffusion.mixed_loss(self, x, cond)
663 x_num_t = self.gaussian_q_sample(x_num, t, noise=noise)
664 if x_cat.shape[1] > 0:
--> 665 log_x_cat = index_to_log_onehot(x_cat.long(), self.num_classes)
666 log_x_cat_t = self.q_sample(log_x_start=log_x_cat, t=t)
668 x_in = torch.cat([x_num_t, log_x_cat_t], dim=1)

File /opt/venv/lib/python3.10/site-packages/synthcity/plugins/core/models/tabular_ddpm/utils.py:145, in index_to_log_onehot(x, num_classes)
143 onehots = []
144 for i in range(len(num_classes)):
--> 145 onehots.append(F.one_hot(x[:, i], num_classes[i]))
146 x_onehot = torch.cat(onehots, dim=1)
147 log_onehot = torch.log(x_onehot.float().clamp(min=1e-30))

RuntimeError: CUDA error: device-side assert triggered
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.