使用kashgari.utils.load_model 加载模型进行标签预测会报错[BUG]

[Gjh9508]

You must follow the issue template and provide as much information as possible. otherwise, this issue will be closed.
请按照 issue 模板要求填写信息。如果没有按照 issue 模板填写，将会忽略并关闭这个 issue

Check List

Thanks for considering to open an issue. Before you submit your issue, please confirm these boxes are checked.

You can post pictures, but if specific text or code is required to reproduce the issue, please provide the text in a plain text format for easy copy/paste.

• [y ] I have searched in existing issues but did not find the same one.
• [y ] I have read the documents

Environment

• OS [e.g. Mac OS, Linux]: Centos(云服务器环境)
• Python Version: 3.6
• requirements.txt:
• 运行环境：tf1.14，kashgari 1.0.0

import pandas as pd
import numpy as np
import kashgari
from kashgari.tasks.classification import BiLSTM_Model
from kashgari.processors import ClassificationProcessor
from kashgari.embeddings import BareEmbedding
import logging

Issue Description

What

model.save('/output/xxx')
kashgari.utils.load_model 
loaded_model.predict(valid_x[:40])

模型在save之前可以正常使用，但是重新加载模型进行标签预测，会报错。报错信息如下：

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-10-8ca42a11a04b> in <module>
----> 1 loaded_model.predict(valid_x[:40])

/usr/local/lib/python3.6/dist-packages/kashgari/tasks/classification/base_model.py in predict(self, x_data, batch_size, multi_label_threshold, debug_info, predict_kwargs)
     78                 pred = pred.argmax(-1)
     79 
---> 80             res = self.embedding.reverse_numerize_label_sequences(pred)
     81             if debug_info:
     82                 logging.info('input: {}'.format(tensor))

/usr/local/lib/python3.6/dist-packages/kashgari/embeddings/base_embedding.py in reverse_numerize_label_sequences(self, sequences, lengths)
    215                                          sequences,
    216                                          lengths=None):
--> 217         return self.processor.reverse_numerize_label_sequences(sequences, lengths=lengths)
    218 
    219     def __repr__(self):

/usr/local/lib/python3.6/dist-packages/kashgari/processors/classification_processor.py in reverse_numerize_label_sequences(self, sequences, **kwargs)
     82     def reverse_numerize_label_sequences(self, sequences, **kwargs):
     83         if self.multi_label:
---> 84             return self.multi_label_binarizer.inverse_transform(sequences)
     85         else:
     86             return [self.idx2label[label] for label in sequences]

AttributeError: 'NoneType' object has no attribute 'inverse_transform'

Reproduce

[The steps to reproduce this issue. What is the URL you were trying to play, where did you put your code, etc.]

Other Comment

[Add anything else here]

[BrikerMan]

谢谢反馈问题，已修复。
可以按照以下方案安装修复分支测试一下，这个分支过两天作为下一版本发布。
麻烦测试完反馈一下结果。

pip uninstall -y kashgari
pip install git+https://github.com/BrikerMan/Kashgari@fix/multi_label_classification_loading

[Gjh9508]

问题已经解决，谢谢！

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[sly123197811]

您好，我又遇到了相同问题，请问还没有作为最终版本修复吗，2020年5月14日11:22:18

[sly123197811]

谢谢反馈问题，已修复。
可以按照以下方案安装修复分支测试一下，这个分支过两天作为下一版本发布。
麻烦测试完反馈一下结果。

pip uninstall -y kashgari
pip install git+https://github.com/BrikerMan/Kashgari@fix/multi_label_classification_loading

您好，我最近clone的kashgari人仍然存在相同的问题，您不是说作为最终版本修复了吗？

[BrikerMan]

@sly123197811 请提供版本号信息和最小复现代码。

[sly123197811]

@sly123197811 请提供版本号信息和最小复现代码。

Python Version: 3.6
运行环境：tf1.14，kashgari-tf 0.5.5，windows10

复现代码：
modelpath = 'models/'
load_model = kashgari.utils.load_model(modelpath)
sample = list(jieba.cut(testdata))
y = load_model.predict([sample])
print(y[0])

问题跟楼主的一样

[BrikerMan]

执行以下命令更新到最新版本。

pip uninstall -y kashgari-tf
pip install kashgari

更新日志：https://kashgari.readthedocs.io/about/release-notes.html