[python-package] Can't continued training on Dataset with SequenceDataset(lgb.Sequence)

[eromoe]

Description

Can't continued training on Dataset with SequenceDataset(lgb.Sequence)

Reproducible example

I create dataset from lgb.Sequence

class PartitionSequence(lgb.Sequence):
    def __init__(self, data:np.ndarray, batch_size=4096):
        self.data = data
        self.batch_size = batch_size

    def __getitem__(self, idx):
        return self.data[idx]

    def __len__(self):
        return len(self.data)


....
....
data.append(PartitionSequence(X.values, batch_size))
dataset = lgb.Dataset(data, label=y, 
                      feature_name=list(X.columns), 
                      weight=weight, 
                      position=dates, 
                      categorical_feature=cat_cols,
                      free_raw_data=free_raw_data)

When I continue trainning with loop

model=None
for train_data in iter_read(....):
        model = lgb.train(
            params,
            train_data,
            init_model=model,
            num_boost_round=num_boost_round,
            keep_training_booster=True,
        )

Get error

Exception has occurred: ValueError       (note: full exception trace is shown but execution is paused at: _run_module_as_main)
Cannot convert data list to numpy array.
  File "C:\envs\quant\Lib\site-packages\lightgbm\basic.py", line 1174, in predict
    data = np.array(data)
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (24,) + inhomogeneous part.

The above exception was the direct cause of the following exception:

  File "C:\envs\quant\Lib\site-packages\lightgbm\basic.py", line 1176, in predict
    raise ValueError('Cannot convert data list to numpy array.') from err
  File "C:\envs\quant\Lib\site-packages\lightgbm\basic.py", line 1972, in _set_init_score_by_predictor
    init_score: Union[np.ndarray, scipy.sparse.spmatrix] = predictor.predict(
  File "C:\envs\quant\Lib\site-packages\lightgbm\basic.py", line 2801, in _set_predictor
    self._set_init_score_by_predictor(
  File "C:\envs\quant\Lib\site-packages\lightgbm\engine.py", line 202, in train
    ._set_predictor(predictor) \
  File "E:\Workspace\github_me\聚宽项目\notebook_exports.py", line 2271, in train_lgb_ensemble
    m = lgb.train(params, ds,
  File "E:\Workspace\github_me\聚宽项目\train_test_model_scroll.py", line 224, in <module>
    model = train_lgb_ensemble(params, train_data,
  File "C:\envs\quant\Lib\runpy.py", line 86, in _run_code
    exec(code, run_globals)
  File "C:\envs\quant\Lib\runpy.py", line 196, in _run_module_as_main (Current frame)
    return _run_code(code, main_globals, None,
ValueError: Cannot convert data list to numpy array.

Debug into find

Apparently can't convert list of lgb.Sequence into np.array .

Environment info

LightGBM version or commit hash:
lightgbm 4.3.0