[求助]TensorFlow2怎么样安装到宝塔

为了能快速了解并处理您的问题，请提供以下基础信息：

面板、插件版本：8.0

系统版本：Centos 7.6

问题描述：如题，TensorFlow2的Seq2seq中文聊天模型

相关截图（日志、错误）：启动我设置了app.py，前端可以进去了但是发消息就提示error calling function

错误日志是'Traceback (most recent call last):

  File "/www/wwwroot/default/bot/b1b9a972ccd8c962a473909b97007eb4_venv/lib/python3.9/site-packages/flask/app.py", line 2190, in wsgi_app

    response = self.full_dispatch_request()

  File "/www/wwwroot/default/bot/b1b9a972ccd8c962a473909b97007eb4_venv/lib/python3.9/site-packages/flask/app.py", line 1486, in full_dispatch_request

    rv = self.handle_user_exception(e)

  File "/www/wwwroot/default/bot/b1b9a972ccd8c962a473909b97007eb4_venv/lib/python3.9/site-packages/flask/app.py", line 1484, in full_dispatch_request

    rv = self.dispatch_request()

  File "/www/wwwroot/default/bot/b1b9a972ccd8c962a473909b97007eb4_venv/lib/python3.9/site-packages/flask/app.py", line 1469, in dispatch_request

    return self.ensure_sync(self.view_functions[rule.endpoint])(**view_args)

  File "/www/wwwroot/default/bot/web/app.py", line 43, in reply

    res_msg = execute.predict(req_msg)

  File "/www/wwwroot/default/bot/execute.py", line 113, in predict

    inputs = tf.keras.preprocessing.sequence.pad_sequences([inputs],maxlen=max_length_inp,padding='post')

  File "/www/wwwroot/default/bot/b1b9a972ccd8c962a473909b97007eb4_venv/lib/python3.9/site-packages/keras/src/utils/data_utils.py", line 1097, in pad_sequences

    sample_shape = np.asarray(x).shape[1:]

ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (12,) + inhomogeneous part.' 这样子的

我应该设置启动app.py还是其他的文件，还是说搭建有问题，训练时可以成功跑起来的