Created by: yocen

欢迎您反馈PaddleHub使用问题，非常感谢您对PaddleHub的贡献！ 在留下您的问题时，辛苦您同步提供如下信息：

• 版本、环境信息 1）PaddleHub和PaddlePaddle版本：PaddleHub1.6.1，PaddlePaddle1.7.1 2）系统环境：AI STUDIO高级版
• 复现信息： 加载finetune好的senta-bilstm模型跑task.predict，首次运行task.predict正常，紧接着更改预测数据再跑task.predict，会报错，必须重启notebook重新加载模型才能再跑预测。 项目地址：https://aistudio.baidu.com/aistudio/projectdetail/311069 错误信息： ---------------------------------------------------------------------------ValueError Traceback (most recent call last) in 6 run_states = cls_task.predict( 7 data=data, ----> 8 accelerate_mode=False) 9 10 results = [run_state.run_results for run_state in run_states] /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddlehub/finetune/task/base_task.py in predict(self, data, load_best_model, return_result, accelerate_mode) 1014 self.init_if_necessary() 1015 if not self.accelerate_mode: -> 1016 run_states = self._run() 1017 else: 1018 if not self._predictor: /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddlehub/finetune/task/base_task.py in _run(self, do_eval) 1072 period_run_states = [] 1073 -> 1074 for run_step, batch in enumerate(data_reader(), start=1): 1075 step_run_state = RunState(len(self.fetch_list)) 1076 step_run_state.run_step = 1 /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/data_feeder.py in reader_creator() 494 item.append(batch) 495 if len(item) == num: --> 496 yield list(self.feed_parallel(item, num)) 497 item = [] 498 if not drop_last and len(item) != 0: /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/data_feeder.py in feed_parallel(self, iterable, num_places) 407 for p, batch in six.moves.zip(places, iterable): 408 self.place = p --> 409 yield self.feed(batch) 410 self.place = place 411 /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/data_feeder.py in feed(self, iterable) 328 for each_name, each_converter in six.moves.zip(self.feed_names, 329 converter): --> 330 ret_dict[each_name] = each_converter.done() 331 return ret_dict 332 /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/data_feeder.py in done(self) 137 138 def done(self): --> 139 arr = numpy.array(self.data, dtype=self.dtype) 140 if self.shape: 141 if len(arr.shape) != len(self.shape): ValueError: invalid literal for int() with base 10: '酒'