[🐛BUG] AttributeError: module 'recbole.model.exlib_recommender.xgboost' has no attribute 'xgboost'

[dexforint]

I try to learn a Context Aware Model with xgboost but I receive next error:

AttributeError: module 'recbole.model.exlib_recommender.xgboost' has no attribute 'xgboost'

[dexforint]

recbole.version == '1.1.1'

[dexforint]

!pip install recbole

from tqdm.notebook import tqdm
from recbole.quick_start import run_recbole
# models = ['LR', 'FM', 'NFM', 'DeepFM', 'xDeepFM', 'AFM', 'FFM', 'FwFM', 'FNN', 'PNN', 
#          'DSSM', 'WideDeep', 'DCN', 'DCN V2', 'AutoInt', 'xgboost', 'lightgbm']
# 'DIN', 'DIEN'
models = ['xgboost', 'lightgbm']

all_results = []
for model in tqdm(models):
    results = run_recbole(model=model, dataset='ml-100k', config_file_list=['test.yaml'])
    results['model'] = model
    print(results)
    all_results.append(results)
   

[dexforint]

I understood your problem: in your docs Xgboost model initializes like:

run_recbole(model='xgboost', dataset='ml-100k')

But it's wrong, change it to:

run_recbole(model='XGBoost', dataset='ml-100k')

[dexforint]

But I have the following error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[11], line 10
      8 all_results = []
      9 for model in tqdm(models):
---> 10     results = run_recbole(model=model, dataset='ml-100k', config_file_list=['test.yaml'])
     11     results['model'] = model
     12     print(results)

File /opt/conda/lib/python3.10/site-packages/recbole/quick_start/quick_start.py:88, in run_recbole(model, dataset, config_file_list, config_dict, saved)
     85 trainer = get_trainer(config["MODEL_TYPE"], config["model"])(config, model)
     87 # model training
---> 88 best_valid_score, best_valid_result = trainer.fit(
     89     train_data, valid_data, saved=saved, show_progress=config["show_progress"]
     90 )
     92 # model evaluation
     93 test_result = trainer.evaluate(
     94     test_data, load_best_model=saved, show_progress=config["show_progress"]
     95 )

File /opt/conda/lib/python3.10/site-packages/recbole/trainer/trainer.py:1005, in DecisionTreeTrainer.fit(self, train_data, valid_data, verbose, saved, show_progress)
   1001 def fit(
   1002     self, train_data, valid_data=None, verbose=True, saved=True, show_progress=False
   1003 ):
   1004     for epoch_idx in range(self.epochs):
-> 1005         self._train_at_once(train_data, valid_data)
   1007         if (epoch_idx + 1) % self.eval_step == 0:
   1008             # evaluate
   1009             valid_start_time = time()

File /opt/conda/lib/python3.10/site-packages/recbole/trainer/trainer.py:1114, in XGBoostTrainer._train_at_once(self, train_data, valid_data)
   1112 self.dvalid = self._interaction_to_lib_datatype(valid_data)
   1113 self.evals = [(self.dtrain, "train"), (self.dvalid, "valid")]
-> 1114 self.model = self.xgb.train(
   1115     self.params,
   1116     self.dtrain,
   1117     self.num_boost_round,
   1118     self.evals,
   1119     early_stopping_rounds=self.early_stopping_rounds,
   1120     evals_result=self.evals_result,
   1121     verbose_eval=self.verbose_eval,
   1122     xgb_model=self.boost_model,
   1123     callbacks=self.callbacks,
   1124 )
   1126 self.model.save_model(self.temp_file)
   1127 self.boost_model = self.temp_file

File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:620, in require_keyword_args.<locals>.throw_if.<locals>.inner_f(*args, **kwargs)
    618 for k, arg in zip(sig.parameters, args):
    619     kwargs[k] = arg
--> 620 return func(**kwargs)

File /opt/conda/lib/python3.10/site-packages/xgboost/training.py:182, in train(params, dtrain, num_boost_round, evals, obj, feval, maximize, early_stopping_rounds, evals_result, verbose_eval, xgb_model, callbacks, custom_metric)
    171 cb_container = CallbackContainer(
    172     callbacks,
    173     metric=metric_fn,
   (...)
    177     output_margin=callable(obj) or metric_fn is feval,
    178 )
    180 bst = cb_container.before_training(bst)
--> 182 for i in range(start_iteration, num_boost_round):
    183     if cb_container.before_iteration(bst, i, dtrain, evals):
    184         break

TypeError: 'NoneType' object cannot be interpreted as an integer

[dexforint]

I solved this problem too: I just added into config:

xgb_num_boost_round: 100

[dexforint]

Please add default value to this parameter

[dexforint]

Same thing with LightGBM, add default values:

lgb_params: {}
lgb_num_boost_round: 300

[Fotiligner]

@dexforint Thank you for your suggestions! We have just added the config information into our docs and config file. And Thanks again for your advises.