Dataset_M4 数据读取有 bug

__read_data__中 training_values = np.array( [v[~np.isnan(v)] for v in dataset.values[dataset.groups == self.seasonal_patterns]]) # split different frequencies 修改为 training_values = np.array( [v[~np.isnan(v)] for v in dataset.values[dataset.groups == self.seasonal_patterns]],dtype=object) # split different frequencies ,因为数据长度不一样